Skip to main content

Foundational Trials

HYPOLYTE

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Hydrocortisone Therapy for Patients With Multiple Trauma: The Randomized Controlled HYPOLYTE Study
  • Acronym: HYPOLYTE
  • Year: 2011
  • Journal published in: JAMA
  • Citation: Roquilly A, Mahe PJ, Seguin P, Guitton C, Floch H, Tellier AC, et al. Hydrocortisone therapy for patients with multiple trauma: the randomized controlled HYPOLYTE study. JAMA. 2011;305(12):1201-1209.

Context & Rationale

  • Background
    • Mechanically ventilated patients with multiple trauma have a high burden of hospital-acquired pneumonia (HAP), which prolongs ventilation and intensive care and is linked to worse clinical outcomes.
    • Post-traumatic inflammation and subsequent immune dysfunction were hypothesised to contribute to infection susceptibility.
    • Critical illness–related corticosteroid insufficiency (CIRCI) had been described in other ICU syndromes (notably septic shock) and was increasingly recognised after major trauma, potentially exacerbated by etomidate exposure.
    • Low-dose (“stress-dose”) hydrocortisone can modulate inflammatory responses and improve haemodynamic stability; whether it could prevent infection after severe trauma was uncertain.
  • Research Question/Hypothesis
    • In adults with multiple trauma expected to require >48 hours of mechanical ventilation, does early protocolised hydrocortisone (continued only when CIRCI is present) reduce time-to-first HAP within 28 days compared with placebo?
  • Why This Matters
    • HAP after trauma is common and clinically consequential; effective prevention could meaningfully reduce ventilator time and ICU utilisation.
    • Hydrocortisone is widely available and inexpensive, but carries plausible harms (metabolic complications, neuromuscular weakness, superinfection); randomised evidence was needed to estimate net benefit in this setting.
    • The trial tests whether treating trauma-associated CIRCI (or etomidate-related adrenal suppression) translates into patient-centred benefits rather than biochemical correction alone.

Design & Methods

  • Research Question: Among intubated adults with multiple trauma enrolled early after injury, does protocolised hydrocortisone (continued for 7 days only in patients with CIRCI) reduce the incidence/time-to-first hospital-acquired pneumonia by day 28 compared with placebo?
  • Study Type:
    • Multicentre, randomised, double-blind, placebo-controlled, parallel-group trial (investigator-initiated).
    • Setting: 7 intensive care units in France (enrolment November 2006 to August 2009).
    • Central randomisation with block size 4; stratified by centre, severe traumatic brain injury (TBI), and injury severity score (ISS) >30.
  • Population:
    • Adults aged >15 years and 3 months with multiple trauma (≥2 injuries and ISS >15), enrolled within 36 hours of trauma onset, expected to require mechanical ventilation >48 hours; all patients were intubated before randomisation.
    • Key exclusions: pre-existing adrenal insufficiency; immunodeficiency/immunosuppression; ongoing systemic corticosteroids; pregnancy; anticipated death within 24 hours; prior inclusion in another trial; refusal/withdrawal of consent.
    • Baseline physiological classification: short corticotropin (250 μg) test to define “corticosteroid insufficiency” (basal cortisol <15 μg/dL or increment <9 μg/dL) vs “adapted corticosteroid function”.
  • Intervention:
    • Hydrocortisone hemisuccinate intravenously: 50 mg bolus then continuous infusion 200 mg/day for 5 days, then 100 mg on day 6 and 50 mg on day 7.
    • Initiated immediately after completion of the baseline corticotropin test and within 36 hours of trauma onset.
    • Stopping rule: if adapted corticosteroid function, study drug discontinued when test results available (median exposure 34 hours; IQR 20–49).
  • Comparison:
    • Matching placebo bolus and continuous infusion with identical tapering and stopping rules, plus usual ICU care.
    • Pneumonia-prevention and ICU co-interventions permitted and recorded (eg, antibiotic prophylaxis, stress ulcer prophylaxis, enteral nutrition, semi-recumbent positioning, glycaemic control).
  • Blinding:
    • Double-blind (patients, clinicians, investigators, and monitoring board blinded); an unblinded pharmacist prepared indistinguishable study syringes.
    • Baseline corticotropin testing guided continuation vs discontinuation of study drug, but treatment allocation (hydrocortisone vs placebo) remained concealed.
  • Statistics:
    • Power calculation: 45 patients per group with corticosteroid insufficiency were required to detect a 20% absolute reduction in HAP (from 50% to 30%) with 80% power at the 5% significance level; assuming 50% prevalence of insufficiency, total target sample size was 180.
    • Planned interim analysis after 75 patients using O’Brien–Fleming boundaries (α=0.005 at interim; α=0.048 at final); observed insufficiency ≈70% led to revised total sample size of 150.
    • Primary analysis: intention-to-treat (ITT) time-to-event analysis for first HAP episode by day 28 using Cox proportional hazards; pre-specified modified ITT analysis in the corticosteroid-insufficient subgroup (CIRCI).
  • Follow-Up Period: 28 days after randomisation (or until death), with daily ICU data collection and predefined outcome assessment windows.

Key Results

This trial was stopped early. Recruitment stopped at 150 patients (vs 180 planned) after the planned interim analysis/sample size reassessment because the observed prevalence of corticosteroid insufficiency was higher than anticipated, meeting the required sample for the powered subgroup analysis.

Outcome Hydrocortisone Placebo Effect p value / 95% CI Notes
Hospital-acquired pneumonia by day 28 (primary; ITT) 26/73 (35.6%) 39/76 (51.3%) HR 0.51 95% CI 0.30 to 0.83; P=0.007 Time-to-first HAP episode within 28 days.
Hospital-acquired pneumonia by day 28 (CIRCI subgroup; modified ITT) 20/56 (35.7%) 31/57 (54.4%) HR 0.47 95% CI 0.25 to 0.86; P=0.01 Pre-specified subgroup defined by basal cortisol <15 μg/dL or increment <9 μg/dL after corticotropin.
Hospital-acquired pneumonia by day 28 (CIRCI + traumatic brain injury subgroup) 13/32 (40.6%) 25/35 (71.4%) HR 0.36 95% CI 0.17 to 0.74; P=0.005 Interaction P=0.12 (no statistically significant heterogeneity by TBI status).
Mechanical ventilation–free days (0–28 days; ITT) 16 (SD 8) 12 (SD 9) Mean difference +4 days 95% CI 2 to 7; P=0.001 Ventilator-free days increased; aligns with shorter duration of ventilation.
Duration of mechanical ventilation (days; ITT) 8 (IQR 4 to 15) 13 (IQR 10 to 22) Median difference −5 days 95% CI −12 to −1; P=0.007 Consistent with earlier successful weaning (time-to-event HR 1.71; 95% CI 1.20 to 2.44; P=0.003).
ICU length of stay (days; ITT) 18 (SD 15) 24 (SD 16) Mean difference −6 days 95% CI −11 to −1; P=0.03 Resource utilisation signal driven by shorter ventilation and fewer pulmonary complications.
Acute lung injury / ARDS (ITT) 3/73 (4.1%) 11/76 (14.5%) Absolute difference −10% 95% CI −19% to −1%; P=0.04 Secondary pulmonary morbidity outcome; small event numbers.
28-day mortality (ITT) 6/73 (8.2%) 4/76 (5.3%) Absolute difference +3% 95% CI −5% to 11%; P=0.44 Trial not powered for mortality; wide confidence interval.
Hyponatraemia (ITT) 0/73 (0%) 7/76 (9.2%) Absolute difference −9% 95% CI −16% to −3%; P=0.01 Metabolic tolerance signal favouring hydrocortisone group.
Gastrointestinal bleeding/perforation (ITT) 1/73 (1.4%) 0/76 (0%) Absolute difference +1% 95% CI −2% to 4%; P=0.49 Rare events; imprecision substantial.
Corticosteroid insufficiency on day 8 (among patients assessed) 25/40 (62.5%) 17/51 (33.3%) Absolute difference +29% 95% CI 9% to 49%; P=0.001 Biochemical endpoint suggesting more persistent adrenal suppression after hydrocortisone exposure; clinical impact not established.
  • Hydrocortisone was associated with a large absolute reduction in HAP (35.6% vs 51.3%) and clinically meaningful shortening of ventilation and ICU stay, without a demonstrated mortality benefit.
  • Benefit was most apparent in the pre-specified CIRCI subgroup (HR 0.47) and numerically strongest in patients with TBI, although interaction testing did not show statistically significant heterogeneity.
  • Adverse events were uncommon, but biochemical follow-up suggested more persistent adrenal suppression at day 8 in the hydrocortisone group (62.5% vs 33.3%).

Internal Validity

  • Randomisation and Allocation:
    • Central computer-generated randomisation; block size 4; stratified by centre, severe TBI, and ISS >30.
    • Allocation concealment supported by pharmacy-controlled preparation; ICU clinicians and investigators remained blinded to treatment assignment.
  • Drop out or exclusions:
    • 150 patients enrolled; 1 excluded after withdrawal of consent; ITT analysis included 149 patients (73 hydrocortisone; 76 placebo).
    • Loss to follow-up for primary endpoint was not reported as a major issue; mortality ascertainment to day 28 appears complete in analysed cohort.
  • Performance/Detection Bias:
    • Double-blind design reduces differential co-interventions and outcome ascertainment bias.
    • Pneumonia diagnosis incorporated objective components (radiographic infiltrate plus quantitative microbiology thresholds), but still relies on clinical sampling decisions and interpretation; central adjudication was not reported.
    • Corticotropin testing determined continuation vs discontinuation of study drug (knowledge of adrenal function status), but this was applied identically in both arms and does not reveal allocation.
  • Protocol Adherence:
    • Treatment initiated early: mean time from trauma to study drug start was 25 (SD 13) hours with hydrocortisone vs 25 (SD 15) hours with placebo.
    • Baseline CIRCI prevalence was balanced: 56/73 (76.7%) hydrocortisone vs 57/76 (75.0%) placebo.
    • Planned discontinuation in adapted corticosteroid function occurred after median 34 hours (IQR 20–49), limiting steroid exposure in that stratum and concentrating biological separation within CIRCI patients.
  • Baseline Characteristics:
    • Groups were broadly comparable: age 36 (SD 15) vs 36 (SD 16); men 56/73 (76.7%) vs 61/76 (80.3%).
    • Trauma burden similar: ISS median 29 (IQR 22–38) vs 27 (IQR 22–36); traumatic brain injury 42/73 (57.5%) vs 42/76 (55.3%).
    • Etomidate exposure was common and balanced: 44/73 (62.0%) vs 50/76 (65.8%).
  • Heterogeneity:
    • Effect estimates differed by physiological stratum: benefit was not apparent in the “adapted corticosteroid function” group, consistent with limited steroid exposure and a biologically enriched treatment effect in CIRCI.
    • TBI subgroup showed a larger effect size (HR 0.36), but interaction testing (P=0.12) does not support definitive effect modification.
  • Timing:
    • Intervention delivered within the hypothesised early window (≤36 hours from trauma), potentially before established nosocomial infection and during early inflammatory perturbation.
  • Dose:
    • Hydrocortisone regimen (200 mg/day with taper over 7 days) mirrors stress-dose strategies used in other critical illness contexts.
    • Whether a shorter course, lower dose, or avoidance of hydrocortisone in non-CIRCI patients would preserve benefit while reducing adrenal suppression is not established within this trial.
  • Separation of the Variable of Interest:
    • Pneumonia-prevention co-interventions were similar: antibiotic prophylaxis 63/73 (86.3%) vs 65/76 (85.5%); oropharyngeal decontamination 44/73 (61.1%) vs 47/76 (61.8%).
    • Other supportive practices were similar: semi-recumbent positioning 62/71 (87.3%) vs 67/76 (88.2%); glycaemic control protocol 68/72 (94.4%) vs 71/75 (94.7%); enteral nutrition by day 7 64/72 (88.9%) vs 69/76 (90.8%).
    • Clinical separation in outcomes was substantial: mechanical ventilation–free days 16 (SD 8) vs 12 (SD 9); ICU length of stay 18 (SD 15) vs 24 (SD 16).
  • Key Delivery Aspects:
    • Patients were enrolled early and were sufficiently high risk (all intubated; expected prolonged ventilation; HAP incidence 51.3% in placebo), creating a setting where prevention effects are measurable.
    • High etomidate use raises aetiological complexity (trauma CIRCI vs drug-induced adrenal suppression), but exposure was balanced at baseline.
  • Outcome Assessment:
    • Primary outcome was time-to-first HAP episode to day 28, based on clinical/radiographic criteria plus quantitative microbiological thresholds.
    • Mortality and major complications were objectively defined but infrequent, limiting precision.
  • Statistical Rigor:
    • Interim monitoring used an O’Brien–Fleming approach (final α=0.048), and the observed primary outcome P value (0.007) remains below this threshold.
    • Trial was powered for pneumonia reduction in the CIRCI subgroup, not mortality; early stopping and modest sample size increase the risk that effect sizes are overestimated.

Conclusion on Internal Validity: Overall, internal validity appears moderate to strong given central concealed randomisation, double blinding, minimal attrition, and broadly balanced co-interventions; the main threats are early stopping/small sample size, dependence on a clinician-diagnosed infection endpoint, and the biological complexity introduced by prevalent etomidate exposure.

External Validity

  • Population Representativeness:
    • Participants were predominantly young adults (mean age 36 years), mostly male, with substantial traumatic brain injury burden (~56%), and all were intubated early with anticipated prolonged ventilation.
    • High etomidate use (~64%) may not reflect current practice in all trauma systems and could influence CIRCI prevalence and treatment responsiveness.
    • Patients with immunosuppression, pregnancy, chronic steroids, and expected early death were excluded, limiting applicability to those groups.
  • Applicability:
    • Intervention requires early enrolment and access to a corticotropin test with rapid operationalisation of stopping rules; feasibility may vary across centres and health systems.
    • Background ICU prevention practices (high rates of prophylactic antibiotics and other bundles) may differ internationally, potentially modifying absolute benefit.
    • Given low mortality and focus on morbidity endpoints, extrapolation to populations with different injury patterns, older age, or higher baseline mortality is uncertain.

Conclusion on External Validity: Generalisability is moderate to similar high-resource ICUs managing early-intubated severe trauma, but is limited by the need for early physiological testing, high etomidate exposure, and uncertainty about applicability to broader trauma cohorts and evolving contemporary ventilation/bundle practices.

Strengths & Limitations

  • Strengths:
    • Randomised, double-blind, placebo-controlled, multicentre ICU trial with central allocation concealment.
    • Early initiation of therapy within a biologically plausible window (≤36 hours post-trauma).
    • Clinically meaningful morbidity endpoints with quantitative microbiological criteria for pneumonia.
    • Prespecified interim monitoring and an a priori physiological subgroup (CIRCI) aligned with mechanistic intent.
  • Limitations:
    • Stopped early with modest sample size; increases risk of an overestimated treatment effect and limits precision for uncommon harms and mortality.
    • Primary endpoint is an infection diagnosis that can be influenced by clinical suspicion and sampling; central adjudication not reported.
    • CIRCI definition and corticotropin testing thresholds are debated and may not translate across contemporary practice frameworks.
    • Very high prevalence of etomidate exposure complicates causal attribution (treatment of etomidate-induced suppression vs trauma-related CIRCI).
    • Biochemical follow-up suggested more persistent adrenal suppression at day 8 after hydrocortisone exposure, with unclear longer-term clinical implications.

Interpretation & Why It Matters

  • Clinical signal
    • Early protocolised hydrocortisone reduced HAP and shortened ventilation and ICU length of stay in a high-risk, intubated trauma ICU cohort.
    • Mortality was low and not improved; the confidence interval is compatible with both modest harm and modest benefit.
  • Targeting and mechanism
    • Benefit clustered in patients meeting the trial’s CIRCI definition, consistent with a “replacement/normalisation” concept rather than broad immunosuppression.
    • Earlier liberation from ventilation may be both a mediator and a competing-risk pathway for reduced pneumonia incidence.
  • Implications for practice
    • Findings support equipoise rather than routine adoption: the evidence base is limited (single small trial with early stopping) and the clinical community’s tolerance for infection-endpoint bias and unmeasured harms varies.
    • If considered, implementation would require careful attention to early timing, physiological selection, and monitoring for endocrine/metabolic complications.

Controversies & Other Evidence

  • Editorial perspective and trial interpretability:
    • The accompanying editorial emphasised that this was a small, early-stopped trial focused on morbidity (pneumonia) rather than mortality, and that the apparent benefit required confirmation in larger studies before changing routine trauma practice.1
  • Correspondence: definition of CIRCI, diagnostic susceptibility, and harms:
    • Concerns were raised that the corticotropin test/CIRCI definition might be unreliable for guiding therapy and that benefits might not be clearly confined to true adrenal dysfunction; authors responded with sensitivity analyses using alternative definitions and argued the effect persisted, particularly in CIRCI.25
    • Diagnostic bias for pneumonia was proposed (eg, steroid antipyretic/anti-inflammatory effects potentially altering clinical thresholds for sampling/diagnosis); authors replied that fever was present in most HAP episodes and diagnostic criteria incorporated quantitative microbiology.35
    • Concerns about underpowered mortality estimates and broader infection harms were highlighted; authors maintained that the study’s key signal was reduced pulmonary infection with shortened ventilation and ICU stay, acknowledging the need for confirmatory work.45
  • Wider trauma/TBI corticosteroid context:
    • High-dose corticosteroids in traumatic brain injury increased mortality in a large pragmatic trial, and contemporary severe TBI guidelines recommend against corticosteroids; this shapes risk perception and constrains extrapolation from low-dose hydrocortisone signals in trauma populations with high TBI prevalence.67
  • Subsequent evidence (confirmatory/adjacent trials):
    • A larger double-blind trial in severe TBI (Corti-TC) tested hydrocortisone plus fludrocortisone and did not significantly reduce HAP (74/165 [45%] vs 87/163 [53%]; HR 0.75; 95% CI 0.55 to 1.03; P=0.07), with underpowering due to lower-than-expected event rates; this attenuates confidence in the magnitude and reproducibility of the HYPOLYTE effect.8
  • Etomidate as confounder or mechanistic modifier:
    • A prespecified analysis from the HYPOLYTE programme found etomidate exposure associated with increased pneumonia susceptibility (HR 2.48; 95% CI 1.19 to 5.18) and suggested hydrocortisone reduced HAP among etomidate-exposed patients (40% vs 62%); this supports biological plausibility but also implies that benefit may partly reflect treatment of etomidate-induced adrenal suppression rather than trauma-specific immunomodulation.9
  • Guidelines and synthesis statements:
    • Society guidance on CIRCI and corticosteroid use suggests against corticosteroids in major trauma (conditional recommendation; low-quality evidence), reflecting that the trauma-specific RCT evidence base remains limited and not definitive for routine implementation.10

Summary

  • In 149 intubated adults with multiple trauma enrolled within 36 hours, early protocolised hydrocortisone reduced hospital-acquired pneumonia by day 28 (35.6% vs 51.3%; HR 0.51; 95% CI 0.30 to 0.83; P=0.007).
  • Hydrocortisone increased ventilator-free days (16 [SD 8] vs 12 [SD 9]; mean difference +4 days; 95% CI 2 to 7; P=0.001) and shortened ICU stay (18 [SD 15] vs 24 [SD 16]; mean difference −6 days; 95% CI −11 to −1; P=0.03).
  • Effects were concentrated in patients meeting the trial’s CIRCI definition, with a larger effect estimate in traumatic brain injury, although statistical interaction was not significant.
  • Mortality was not reduced (8.2% vs 5.3%; P=0.44) and harms were imprecisely estimated; biochemical follow-up suggested more frequent corticosteroid insufficiency at day 8 after hydrocortisone (62.5% vs 33.3%; P=0.001).
  • Subsequent evidence in severe TBI (Corti-TC) did not confirm a statistically significant pneumonia reduction, and guidelines have not endorsed routine steroid prophylaxis in major trauma.

Overall Takeaway

HYPOLYTE is an important early mechanistic RCT suggesting that stress-dose hydrocortisone, when started within 36 hours of severe trauma and continued in patients with CIRCI, can reduce hospital-acquired pneumonia and shorten ventilation and ICU stay. However, the trial’s early stopping, modest sample size, and subsequent mixed confirmatory evidence mean its findings remain hypothesis-generating rather than practice-changing, and routine use in major trauma is not guideline supported.

Overall Summary

  • Early protocolised hydrocortisone in intubated severe trauma reduced HAP and shortened ICU stay, but evidence remains limited and not definitive for routine adoption.
  • High etomidate exposure and debated CIRCI definitions complicate interpretation and may influence who (if anyone) benefits most.

Bibliography