Skip to main content

Foundational Trials

REACT2

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Immediate total-body CT scanning versus conventional imaging and selective CT scanning in patients with severe trauma (REACT-2): a randomised controlled trial
  • Acronym: REACT-2
  • Year: 2016
  • Journal published in: The Lancet
  • Citation: Sierink JC, Treskes K, Edwards MJR, Beuker BJA, den Hartog D, Hohmann J, et al; REACT-2 study group. Immediate total-body CT scanning versus conventional imaging and selective CT scanning in patients with severe trauma (REACT-2): a randomised controlled trial. Lancet. 2016;388(10045):673-683.

Context & Rationale

  • Background
    • Whole-body CT (“pan-scan”) during initial trauma resuscitation became common in high-resource trauma systems, driven by diagnostic completeness and the premise that earlier definitive injury mapping improves triage to surgery, interventional radiology, or ICU.
    • Pre-REACT-2 evidence was dominated by retrospective registry studies and systematic reviews, with strong confounding by indication and centre-level co-interventions (eg, trauma system maturity, staffing, parallel processing). 1
    • Potential harms included higher radiation dose, contrast exposure, resource use, and opportunity costs in the resuscitation pathway, motivating an RCT focused on patient-centred outcomes (mortality) rather than diagnostic yield alone.
  • Research Question/Hypothesis
    • Does immediate total-body CT (iTBCT) as the default imaging strategy during initial evaluation of adults with suspected severe trauma reduce in-hospital mortality compared with conventional radiography/FAST plus selective CT?
    • Hypothesis: iTBCT would reduce mortality by accelerating diagnosis and definitive management, without unacceptable harms.
  • Why This Matters
    • Trauma deaths are time-sensitive; even small, system-level time gains might translate into lives saved if they reliably accelerate haemorrhage control or neurocritical interventions.
    • Because iTBCT consumes scanner, staff, and workflow capacity, an RCT was needed to justify (or refute) routine “pan-scan for severe trauma” as a default strategy rather than a selective tool.

Design & Methods

  • Research Question: In adults with suspected severe trauma, does immediate total-body CT during initial evaluation reduce in-hospital mortality compared with conventional imaging and selective CT?
  • Study Type: Multicentre, randomised, controlled, parallel-group, open-label trial conducted in the emergency department/trauma resuscitation setting at five level-1 trauma centres (four Netherlands; one Switzerland).
  • Population:
    • Setting: Trauma resuscitation room / emergency department during initial evaluation immediately after arrival.
    • Inclusion (non-pregnant adults): ≥18 years with suspected severe injury meeting pre-defined criteria (compromised vital parameters, clinical suspicion of specific severe injuries, and/or high-risk trauma mechanism). 2
    • Key inclusion thresholds (examples): systolic BP <100 mm Hg; heart rate >120 beats/min; GCS <14; respiratory rate >30 or <10 breaths/min; or clinical suspicion of major thoracic/abdominal/pelvic injury or multiple long-bone fractures; or high-risk mechanisms (eg, fall >3 m, ejection, run over/crush).
    • Key exclusions (examples): age <18 years; pregnancy; inter-hospital transfer; low-energy blunt trauma; isolated penetrating injury; and patients requiring immediate CPR or immediate operative intervention with imminent death (too unstable to randomise safely).
  • Intervention:
    • Immediate total-body CT (iTBCT): CT from head to pelvis as part of the initial diagnostic strategy after the primary survey, using a standardised two-step protocol (vertex to pubic symphysis; 64-slice scanners; intravenous contrast where appropriate).
    • Operational intent: Replace the conventional sequence (radiography/FAST then selective CT) with an “early comprehensive” CT-first pathway to expedite diagnosis and downstream management.
  • Comparison:
    • Standard work-up: conventional radiological imaging with focused assessment with sonography for trauma (FAST) and/or radiography, followed by selective CT of regions based on pre-specified criteria and clinician judgement.
    • Rescue/co-interventions: treating teams could order additional imaging or interventions as clinically indicated in either arm.
  • Blinding: Unblinded (patients and clinicians not masked); outcomes included objective endpoints (mortality) and process endpoints (times), the latter potentially sensitive to performance effects.
  • Statistics: 539 patients per group were required to detect a 5% absolute reduction in in-hospital mortality (assumed 12% to 7%) with 80% power at a two-sided 5% significance level; primary analysis described as intention-to-treat (with exclusions after randomisation for consent/eligibility as per trial procedures).
  • Follow-Up Period: Index admission (primary endpoint: in-hospital mortality); prespecified follow-up included 30-day outcomes and 6-month readmission/cost outcomes, with additional longer-term patient-reported follow-up in subsequent publications.

Key Results

This trial was not stopped early. Preplanned unmasked interim safety analyses were undertaken during recruitment; enrolment continued to achieve the target analysed sample size.

Outcome Immediate total-body CT Standard work-up Effect p value / 95% CI Notes
In-hospital mortality (primary) 86/541 (16%) 85/542 (16%) Not reported P=0.92 Primary endpoint; analysed in allocated groups.
24-hour mortality 20/541 (4%) 20/542 (4%) Not reported P=0.98 Early mortality; no difference.
30-day mortality (subset with 30-day data) 69/487 (14%) 65/497 (13%) Not reported P=0.65 Subset excludes trauma-room deaths and other prespecified exclusions from this analysis.
In-hospital mortality — polytrauma subgroup 81/362 (22%) 82/331 (25%) Not reported P=0.46 Prespecified subgroup; no significant interaction reported.
In-hospital mortality — traumatic brain injury subgroup 68/178 (38%) 66/151 (44%) Not reported P=0.31 Prespecified subgroup; signal towards lower mortality but not significant.
Time in trauma room (min), median (IQR) 63 (50–79) 72 (55–87) Not reported P<0.0001 Process metric; shorter with iTBCT.
Time to end of imaging (min), median (IQR) 30 (23–38) 37 (28–49) Not reported P<0.0001 Primary diagnostic sequence completed earlier with iTBCT.
Time to diagnosis (min), median (IQR) 50 (38–70) 58 (43–78) Not reported P<0.0001 Earlier “diagnostic completion” with iTBCT.
Radiation exposure in trauma room (mSv), median (IQR) 20.9 (20.6–20.9) 20.6 (9.9–20.8) Not reported P<0.0001 More patients in standard work-up received lower doses; iTBCT increased early exposure distribution.
Total radiation during hospital admission (mSv), median (IQR) 21.0 (20.6–22.9) 20.6 (9.9–22.8) Not reported P<0.0001 Total exposure slightly higher with iTBCT (distributional shift).
Serious adverse events related to CT scanning 3/541 (1%) 1/542 (<1%) Not reported Not reported Five serious events reported overall; all resulted in death; one occurred in a patient excluded after random allocation.
Readmission within 6 months 90/541 (17%) 57/542 (11%) Not reported P=0.01 Higher readmission rate in iTBCT group; mechanism not fully explained in main report.
Total hospital costs within 6 months (€), mean (95% CI) 25,473 (23,589–27,357) 26,014 (24,010–28,019) Not reported P=0.44 No significant cost difference in the primary economic summary reported in the index paper.
  • iTBCT shortened key diagnostic process times (eg, time to end of imaging 30 [23–38] min vs 37 [28–49] min; time to diagnosis 50 [38–70] min vs 58 [43–78] min; both P<0.0001), but this did not translate into lower in-hospital mortality (16% vs 16%; P=0.92).
  • Radiation exposure was higher in the iTBCT pathway (trauma room: 20.9 [20.6–20.9] mSv vs 20.6 [9.9–20.8] mSv; P<0.0001), reflecting an early CT-heavy diagnostic strategy.
  • Readmission within 6 months occurred more often after iTBCT (17% vs 11%; P=0.01), raising questions about downstream utilisation and unintended consequences.

Internal Validity

  • Randomisation and Allocation: 1:1 randomisation using a computerised system (ALEA) after primary survey in the trauma room; centre-level trauma systems were mature, reducing variability from “learning curve” effects.
  • Dropout/exclusions after randomisation: 1403 patients randomised, but primary analysis included 541/702 (77%) in iTBCT and 542/701 (77%) in standard work-up, with exclusions after random allocation (eg, declined participation; incorrect inclusion), creating a modified intention-to-treat population.
  • Performance/Detection Bias: Open-label design; mortality is objective, but process outcomes (timing, imaging completion) are potentially susceptible to workflow and measurement effects.
  • Protocol adherence and crossover: Crossovers occurred (reported 6 in iTBCT vs 18 in standard work-up); overall protocol violations were reported (approximately 9%), and a post-hoc per-protocol analysis excluding crossovers did not change conclusions.
  • Baseline characteristics: Groups were broadly comparable (median age 42 [27–59] vs 45 [26–59]; 76% male in both; median ISS 20 [10–29] vs 19 [9–29]; hypotension at admission 7% vs 8%). Polytrauma and TBI proportions differed modestly (polytrauma 67% vs 61%; TBI 33% vs 28%).
  • Heterogeneity: Multicentre design across five trauma centres supports robustness; however, centre workflows and CT logistics could influence the magnitude of time-to-diagnosis effects.
  • Timing: Randomisation and imaging occurred early during trauma evaluation, but the observed median time savings were in the order of minutes (eg, trauma-room time 63 [50–79] vs 72 [55–87] min), which may be insufficient to affect mortality unless it reliably accelerates time-critical interventions.
  • Dose (intervention fidelity): The iTBCT protocol (standardised two-step acquisition) achieved diagnostic acceleration at the cost of increased early radiation exposure.
  • Separation of the Variable of Interest: Measurable pathway separation was demonstrated (time to end of imaging 30 [23–38] vs 37 [28–49] min; time to diagnosis 50 [38–70] vs 58 [43–78] min; trauma-room time 63 [50–79] vs 72 [55–87] min; all P<0.0001).
  • Outcome Assessment: Primary outcome (in-hospital mortality) is clear and objective; secondary outcomes included process measures, radiation, costs, and readmission.
  • Statistical Rigor: A priori sample size was achieved for the analysed cohort; interim safety monitoring and multiple imputation sensitivity analyses were described, with consistent conclusions.

Conclusion on Internal Validity: Overall, internal validity appears moderate: randomisation and objective primary outcome support causal inference, but substantial post-randomisation exclusions and open-label process measurement introduce potential selection and performance biases.

External Validity

  • Population Representativeness: Predominantly blunt trauma (98–99%), median ISS ~20 with high proportions of polytrauma and TBI, reflecting typical European major trauma case-mix in level-1 centres; isolated penetrating trauma was uncommon and largely excluded.
  • Applicability: Generalisability is strongest for high-resource trauma systems with immediate CT availability and established trauma team workflows; applicability is limited in centres with constrained CT access, different prehospital triage, higher penetrating trauma burden, or less capacity for parallel resuscitation during CT acquisition.
  • System dependency: The clinical value of iTBCT is likely contingent on whether CT is co-located and whether haemorrhage control pathways (OR/IR) can be activated without delay based on CT findings.

Conclusion on External Validity: External validity is moderate: findings translate well to mature, high-resource blunt-trauma systems, but may not extrapolate to settings where CT availability, geography, or trauma epidemiology differ substantially.

Strengths & Limitations

  • Strengths: Pragmatic multicentre RCT in real-world major trauma resuscitation; objective primary endpoint; standardised CT protocol; clinically relevant process and resource outcomes (radiation, costs, readmission).
  • Limitations: Open-label design; substantial post-randomisation exclusions (consent/eligibility) yielding a modified ITT cohort; modest absolute time gains; limited power to detect smaller mortality effects or to robustly support subgroup claims; limited applicability to penetrating trauma and lower-resource systems.

Interpretation & Why It Matters

  • Clinical practice
    Routine iTBCT for broadly defined “suspected severe trauma” improved diagnostic speed but did not improve survival; this supports a selective rather than universal pan-scan strategy in most trauma systems.
  • Mechanistic insight
    The mortality-neutral result despite faster diagnosis implies that, in a mature trauma system, the limiting step for survival may not be diagnostic completeness, but rather definitive haemorrhage control, neuroprotection, and organisational latency that CT alone cannot resolve.
  • Safety and downstream effects
    Higher early radiation exposure and higher 6-month readmission in the iTBCT group emphasise that “faster and more complete imaging” is not unambiguously beneficial and may shift downstream utilisation in unexpected ways.

Controversies & Other Evidence

  • Time savings versus clinically meaningful delay: Commentaries argued that the key question is whether iTBCT changes time-to-haemorrhage control or other time-critical interventions; modest median diagnostic time differences (minutes) might be insufficient to improve survival unless they reliably alter definitive management pathways. 3
  • Selection after randomisation (consent/eligibility): Correspondence highlighted that post-randomisation exclusions and the practicalities of consent in emergency trauma research can distort comparability and threaten the “pure” intention-to-treat effect estimate, especially if excluded patients differ systematically in prognosis or pathway constraints. 45
  • Confounded observational benefit versus RCT neutrality: REACT-2 challenged earlier registry-based survival associations for whole-body CT, which may reflect confounding and centre-level co-interventions rather than the imaging strategy itself. 1
  • Meta-analytic synthesis post-REACT-2: More recent systematic reviews/meta-analyses integrating REACT-2 generally report reduced time in ED/diagnostic completion with whole-body CT, without consistent mortality benefit, and with trade-offs including radiation exposure and resource utilisation. 67
  • Refining who should receive iTBCT: Secondary analyses from the REACT-2 programme aimed to reduce the number of inclusion criteria (from 15 to 10) to better target severely injured patients and reduce radiation exposure in less severely injured patients, reinforcing a “selective iTBCT” framing. 8
  • Guideline positioning: Imaging guidance for major blunt trauma supports CT as central to evaluation, but emphasises appropriateness and selection rather than an unconditional “pan-scan for all”, aligning with the REACT-2 signal that routine iTBCT is not universally outcome-improving. 9
  • Health economics: A later REACT-2 economic evaluation reported that iTBCT can be cost-effective in specific subgroups (eg, multiple trauma or TBI) from a hospital provider perspective, highlighting that “value” may differ by case-mix even when overall mortality is unchanged. 10

Summary

  • In 1083 analysed adults with suspected severe trauma, immediate total-body CT did not reduce in-hospital mortality versus conventional imaging plus selective CT (16% vs 16%; P=0.92).
  • iTBCT accelerated diagnostic processes (time to end of imaging 30 vs 37 min; time to diagnosis 50 vs 58 min; both P<0.0001) but without demonstrable survival benefit.
  • Radiation exposure was higher with iTBCT (trauma room and total admission; both P<0.0001), consistent with a CT-first pathway.
  • Readmission within 6 months was higher after iTBCT (17% vs 11%; P=0.01), while mean 6-month hospital costs did not differ significantly (P=0.44).
  • The trial reframed “pan-scan” as a selective tool: downstream work emphasised refining criteria to target patients most likely to benefit and to minimise avoidable exposure and utilisation.

Overall Takeaway

REACT-2 is a landmark pragmatic RCT showing that, in mature trauma systems, routine immediate total-body CT for broadly selected severe trauma patients improves speed of diagnosis but does not improve survival. Its enduring contribution is shifting the field from assumption-driven “pan-scan for all” towards evidence-informed selectivity—balancing modest process gains against radiation, downstream utilisation, and system constraints.

Overall Summary

  • Immediate total-body CT: faster diagnosis, no mortality benefit (16% vs 16%).
  • Trade-offs: higher radiation exposure and higher 6-month readmission (17% vs 11%).
  • Modern framing: selective iTBCT for the right patient in the right system, not a universal default.

Bibliography