By Simon Carley / November 18, 2026 / Emergency Medicine
For a procedure that is actually quite rare, it’s interesting to wonder how resuscitative thoracotomy (RT) commands such interest in social media, research and publishing. It’s something that is certainly true of St Emlyn’s when I look back on the number of blogs on the topic, and presentations we have done at conferences. There are many reasons why of course, but for me it that it is one of those high-acuity, low-occurrence (HALO) procedures that really challenges all aspects of being an emergency physician, uncertainty, judgement, team leading, technical skills, rarity, high stress, high stakes…… I could go on. So, whilst it is a rare procedure I think there are reasons why so many emergency physicians have such a disproportionate interest in the topic.
Although the technical aspects of the procedure can be taught, rehearsed and assessed, deciding when to perform it is arguably more difficult. The patients are critically injured, the available information is often incomplete, and the decision must usually be made within a very short period of time. (If you want to know more about goldilocks decision making, read this blog here.) We are also dealing with an intervention that most clinicians will perform infrequently, even in relatively busy trauma systems, making it difficult to develop the experience needed to refine patient selection through personal practice alone.
There is substantial literature describing outcomes following resuscitative thoracotomy, but much of it comes from retrospective series, often involving individual centres with particular expertise or selected groups of patients who survived long enough to reach hospital. This creates obvious difficulties when applying the findings to prehospital critical care, where many patients deteriorate or die before reaching a major trauma centre.
The recently published TETRiS study gives us a prospective national evaluation of RT across the UK, incorporating both prehospital and emergency department practice. Although it cannot establish whether thoracotomy improves survival, it does tell us about the patients undergoing the procedure, the pathology encountered and the factors associated with successful outcomes.
The abstract is below, but as always read the paper yourself.
Abstract
Background
Resuscitative thoracotomy (RT) is performed in patients with traumatic cardiac arrest or peri-arrest to treat potentially reversible causes of circulatory collapse. Optimal patient selection and timing remain uncertain, and existing evidence is largely retrospective, often excluding prehospital deaths. This prospective service evaluation examined contemporary RT practice and outcomes across a national major trauma system, including prehospital and emergency department settings.
Methods
Prospective 12-month observational study across 20 UK Helicopter Emergency Medical Services and 26 Major Trauma Centre Emergency Departments, using the PHOTON, TERN, and NaTRIC research networks.
Results
A total of 116 patients underwent RT, of which 82 (70.7%) were performed prehospital. Median age was 30 years, 97 patients (83.6%) were male, and 81 (69.8%) sustained penetrating injury. Suspected cardiac tamponade was the most common indication for RT (72/116 [62.1%]), although operative findings demonstrated exsanguinating haemorrhage in 82 patients (70.7%). Overall, 50 patients (43.1%) achieved return of spontaneous circulation and eight (6.9%) survived to hospital discharge. Survival was strongly associated with pathophysiological cause, duration of arrest, and physiological state at the time of intervention, with substantially better outcomes observed in patients with cardiac tamponade than exsanguination. Patients undergoing emergency department RT demonstrated shorter arrest durations, more favourable physiology, and higher survival than those undergoing prehospital RT. There were no survivors beyond 10 min of traumatic cardiac arrest.
Conclusion
In this prospective national evaluation of RT practice, outcomes were strongly associated with the pathophysiological cause and duration of arrest, and the residual physiology present at the time of intervention. The most favourable outcomes were observed in patients with cardiac tamponade and those undergoing RT before irreversible physiological deterioration.
Ramage L, Tucker H, Marsden MER, et al. Traumatic Emergency Thoracotomy for Resuscitation in Shock (TETRiS): a national service evaluation of resuscitative thoracotomy practice outside the operating theatre. Resuscitation. 2026;229:111371. https://doi.org/10.1016/j.resuscitation.2026.111371
What did the investigators do?
This is a prospective observational service evaluation over 12 months, between February 2022 and January 2023, including consecutive adult and paediatric patients undergoing resuscitative thoracotomy in the prehospital environment or an emergency department. Procedures performed in operating theatres and those undertaken for non-traumatic indications were excluded.
All 29 UK major trauma centres and 21 helicopter emergency medical services (HEMS) were invited to participate, with data ultimately submitted by 26 major trauma centres and 20 HEMS services. This represents 90% and 95% of the respective services and provides an unusually comprehensive picture of contemporary UK practice. The UK has a significant advantage over some other health economies (you know who you are) as a National Health Service (and a cooperative and collaborative air ambulance network) means that we can do these kind of studies. Both the services I work for in Virchester contributed data to the study.
Decisions about whether to undertake thoracotomy, together with the technical aspects of the procedure, remained dependent on local protocols and clinician judgement.
The investigators collected information about injury mechanism, physiological state, indications for thoracotomy, timing, operative findings and clinical outcomes. The primary outcome was survival to hospital discharge, with secondary outcomes including return of spontaneous circulation, survival beyond 24 hours and neurological outcome among survivors.
Remember that this observational data, there was no control group, and the study did not capture all patients in whom thoracotomy might have been considered but was not performed. Consequently, it can describe the characteristics and outcomes of patients undergoing thoracotomy, but it cannot tell us whether those patients would have survived or died without the intervention. As an evidence level it’s quite low, but probably the best we can get at this time.
What did they find?
The study found 116 RT patients, with 82 procedures (71%) performed prehospital and 34 (29%) in emergency departments. The median age was 30 years, approximately 84% were male, and 70% had sustained penetrating injuries. That fits with my experience here in Virchester in terms of the male:female mix but I was surprised at the number of thoracotomies for blunt trauma.
At the time of thoracotomy, 103 patients (89%) were in established traumatic cardiac arrest, while 13 were described as peri-arrest. Overall, 50 patients (43%) achieved return of spontaneous circulation, 16 (14%) survived beyond 24 hours and eight (6.9%) survived to hospital discharge.
Neurological outcome information was available for seven of the eight survivors. Six achieved a Glasgow Outcome Score of 5, indicating a favourable neurological outcome, while one survived with severe neurological disability.
The overall survival figure is broadly consistent with previous studies, although it conceals considerable variation between different patient groups. This variation is probably the most useful aspect of the paper, particularly when considering how the findings might influence prehospital decision-making.
The underlying pathology and why it matters
Suspected cardiac tamponade was the primary indication for thoracotomy in 72 patients (62%). However, tamponade was identified in only 30 patients (26%), including those who also had significant haemorrhage. Exsanguinating haemorrhage was identified in 82 patients (71%), making it the predominant finding in the cohort. This discrepancy is clinically important because the mechanisms of circulatory collapse in tamponade and exsanguination are fundamentally different, even though both may present with profound shock or traumatic cardiac arrest.
In cardiac tamponade, the heart may retain sufficient circulating blood volume and myocardial function to generate an effective circulation once the obstruction to filling is relieved. In this situation, thoracotomy can provide a relatively direct solution to the underlying problem. By contrast, patients with profound exsanguination may have little circulating volume remaining. Although thoracotomy can facilitate internal haemorrhage control, cardiac massage or aortic occlusion, these interventions cannot independently replace lost blood or provide definitive treatment for all sources of haemorrhage.
This fits with what we have been saying for the last year or so (largely based on the LAA data) that we need different strategies for likely exsanguination vs. tamponade. And we now have data from LAA that the location of the penetrating injury can help us determine the probability of either/both.
Three of 13 patients with isolated tamponade survived to discharge (23.1%), compared with two of 18 patients with combined tamponade and exsanguination (11.1%) and three of 82 patients with exsanguination alone (3.7%). These subgroup numbers are small and should not be interpreted as precise estimates of survival probability. Nevertheless, the findings are physiologically plausible and consistent with previous evidence suggesting that thoracotomy is most successful when it rapidly reverses a discrete, treatable cause of circulatory collapse.
For clinicians making these decisions, the challenge is that the underlying pathology is not always apparent before intervention. A patient with penetrating thoracic trauma and rapidly deteriorating physiology may have tamponade, major intrathoracic haemorrhage, or both. The clinical presentation can be similar, particularly when assessment is limited by the urgency of the situation. The LAA paper on location really helps with this, but there is still plenty of uncertainty.
To date I’m not a big fan of cardiac ultrasound to guide decision making in thoracotomy, as it is such a time-critical procedure in TCA. Faffing around with a probe may delay that decision. Also, you can miss pathology on USS, including cardiac injury so I’ve not been keen (although others are increasingly using it). In the study it was used in 18 patients (16%). I. can see its use in the peri-arrest patient, and maybe I need to think again to additionally try and differentiate tamponade vs. exsanguination, but I’m still on the fence with this. In the study we don’t know why USS was/was not used, and there will undoubtedly have been patients in whom immediate thoracotomy was appropriate without further investigation.
However, in patients where there is time to POCUS without delaying treatment, it could identify tamponade from other causes of circulatory collapse and prompt earlier thoracotomy although whether that would improve patient selection or outcomes remains uncertain. Perhaps its bigger role is in identifying cardiac standstill, and hence futility.
Timing, physiological deterioration and the opportunity for successful intervention
The relationship between the duration of traumatic cardiac arrest and survival is something we have seen in other studies and is something that has influenced my practice. Among patients in established cardiac arrest, the median duration of arrest before thoracotomy was one minute in survivors compared with ten minutes in non-survivors. There were no survivors when the reported duration of arrest exceeded ten minutes. I’d potentially take this further and argue that the optimal time is before they have arrested, but I appreciate that is controversial and harder to judge.
Four of the 13 peri-arrest patients survived, as did three of the 13 patients who underwent thoracotomy within one minute of cardiac arrest. There were no survivors among the 15 patients treated between two and five minutes, one survivor among 22 treated between six and ten minutes, and none among the 44 patients treated after more than ten minutes. Literally seconds count…….
Although the distribution is not linear, survival was overwhelmingly concentrated among patients who had not yet arrested or who underwent intervention very soon after arrest.
This is clinically important, as it’s clear that speed matters. However, at the other end we have to be cautious about a ten-minute absolute threshold for futility on this data. The study includes only eight survivors, and the absence of survival beyond a particular duration does not establish that survival is impossible. Also, the precise timing of traumatic cardiac arrest is often uncertain, especially in the prehospital environment where patients may deteriorate progressively through profound shock before pulselessness is recognised. Judging low flow (and how low) vs. no flow is almost impossible without an arterial line or POCUS, and we don’t usually have those.
Arrest duration is also closely related to other factors associated with survival, including witnessed arrest, residual cardiac activity and the presence of signs of life. Patients who arrest in front of a clinical team may receive an intervention almost immediately, while those found in established arrest may have experienced a much longer period without effective circulation. So, I think it’s safe to say that time matters, but rigid time rules are much trickier.
For peri-arrest patients four of 13 patients (30.8%) undergoing thoracotomy before established cardiac arrest survived, compared with four of 103 patients (3.9%) who were already in traumatic cardiac arrest.
We can’t conclude that RT before arrest caused this difference, because the patients were selected according to clinical circumstances and may have differed substantially in their underlying injuries and physiological reserve. However, this makes sense from a pathophysiological perspective and also from my personal experience. My last thoracotomy survivor (completely neurologically intact) was peri-arrest, clearly shocked, getting worse and basically dying, and so it really didn’t seem sensible to wait until they were completely in arrest.
Signs of life and residual physiology
All eight survivors had documented signs of life at the time of thoracotomy, while none of the 61 patients without signs of life survived. Similarly, all eight survivors had cardiac electrical activity rather than asystole, and survival was substantially more common when arrest had been witnessed. Signs of life were defined broadly, incorporating spontaneous movement, respiratory effort, reactive pupils, measurable blood pressure, palpable pulse, organised cardiac electrical activity or cardiac contractility on ultrasound. Consequently, a patient could be classified as having signs of life without demonstrating obvious movement or respiratory effort.
This matters because the various markers of residual physiology are not independent. Organised cardiac electrical activity is itself included in the definition of signs of life, while witnessed arrest and short arrest duration are also closely related to the likelihood of retaining some physiological activity. However, the numbers are small and so multivariable analyses could not be done, so we can’t statistically determine which of these factors independently predicts survival.
Prehospital versus emergency department thoracotomy
Survival following emergency department thoracotomy was 14.7%, compared with 3.7% following prehospital thoracotomy. The patients undergoing emergency department thoracotomy were more likely to be peri-arrest, to have a witnessed arrest and to demonstrate signs of life. Most importantly, their duration of cardiac arrest before thoracotomy was generally much shorter. Although the median interval from injury to thoracotomy was 51 minutes in the emergency department group compared with 31 minutes prehospital, the median duration of cardiac arrest was only two minutes in the ED group compared with ten minutes prehospital. Again – time seems to really matter.
This paradox can also be explained by the different clinical trajectories. Patients who survive long enough to reach a major trauma centre may arrest later in their course, often with trained clinicians and appropriate equipment immediately available. Those who arrest at the scene may experience a substantially longer period of absent circulation before a thoracotomy can be performed, even though the procedure occurs sooner after the original injury. A faster deterioration also implies more substantial injury patterns (though we don’t know this from the data).
The difference is likely to be biased by patient selection, timing and physiological condition. Also, the study does not support the idea that moving thoracotomy from prehospital to ED improves survival, and pathophysiologically that would make no sense at all.
How much confidence should we place in the findings?
The authors did get a pretty good coverage across the UK (90% of UK major trauma centres and 95% of HEMS services). Including PHEM was also an important finding, but that only covers some PHEM. For example, one of my thoracotomy survivors was done with BASICS and so would not have featured in this data. There may well be other cases missed.
However, there are limitations.
The first is that the study includes only patients who actually underwent thoracotomy. We do not know how many patients were considered for the procedure but did not receive it, or how many might have survived with alternative treatment. Without that information, we cannot assess whether current selection criteria are too restrictive or too permissive, nor can we estimate the effectiveness of thoracotomy itself.
The second limitation is the small number of survivors. Although 116 procedures represent a fairly large cohort for such a rare intervention, only eight patients survived to discharge. This limits statistical precision and makes subgroup comparisons vulnerable.
Although several associations reached statistical significance these should be viewed as exploratory rather than definitive. The number of events was insufficient for robust adjustment for confounding, and the study cannot establish that arrest duration, procedural setting or physiological markers independently caused differences in survival.
There is also potential for misclassification of the underlying pathology, particularly in patients with combined tamponade and exsanguination. The distinction between peri-arrest and established traumatic cardiac arrest may be similarly difficult to make consistently across services.
Finally, local protocols, clinician experience and available resources varied between participating organisations. This heterogeneity makes the study representative of real-world practice but complicates comparisons between groups.
These limitations should not devalue what the authors have done. This is a really important study, but the nature of the study design inevitably has limitations.
What does this mean for UK emergency and prehospital practice?
I don’t think TETRiS provides enough evidence to change thoracotomy indications solely on the basis of its findings. However, it does fit with several other studies and much of what we have said here on St Emlyn’s.
The first is the importance of identifying the dominant reversible pathology. Although thoracotomy can provide several potential interventions, its likelihood of success depends on whether those interventions address the principal cause of circulatory collapse. A patient with isolated tamponade is different to one who has exsanguinated and the priorities of treatment must reflect that distinction.
The second is the importance of recognising the patient’s physiological trajectory. Trajectory is a word I use a lot when teaching resuscitation principles, and here it really makes a difference. Arrest duration, witnessed collapse, signs of life and cardiac electrical activity all appear to be associated with survival. They are closely related and should not be treated as independent predictors. Their combined interpretation may be more useful than applying any single criterion in isolation.
The third is that the decision to undertake thoracotomy must be considered alongside other interventions, particularly haemorrhage control, blood-product resuscitation and rapid access to definitive surgical care. In patients with profound exsanguination, an intervention that delays these priorities may be counterproductive, even if it is technically successful. What to do first, and then what next etc. is really poorly defined at the moment and comes down to a bit of judgement (though a big subclavian features in my approach).
It would be great if this study followed up with an ongoing registry by the way…… if we could learn from this and take it forward that would be great.
The relatively low number of procedures undertaken by most participating services also highlights the importance of maintaining competence in HALO interventions. The median annual volume was only two procedures per HEMS service and one per major trauma centre, with considerable geographical variation. The variation will be huge though. In Virchester it’s a lot higher than that (but that also means it’s probably much less elsewhere).
Training, simulation, team preparation and structured review of cases will therefore remain essential, even in services with substantial trauma experience.
The bottom line
TETRiS is an important contribution to the UK resuscitative thoracotomy literature. The overall survival rate of 6.9% is probably higher than some think. Patients with isolated tamponade, residual physiological activity and short durations of cardiac arrest had substantially better outcomes than those with profound exsanguination or prolonged established arrest.
Successful resuscitation depends on more than technical proficiency. We need to recognise the right pathology, identify the patient who retains a realistic opportunity for recovery and intervene before that opportunity disappears. Decisions, judgement, training and experience are clearly as important as the procedure itself.
vb
S (with input from Zaf Qasim)
This article is an independent St Emlyn’s critical appraisal. Clinical decisions should follow appropriate local protocols, trauma-network arrangements and specialist judgement.
References and further reading
- Ramage L, Tucker H, Marsden MER, et al. Traumatic Emergency Thoracotomy for Resuscitation in Shock (TETRiS): a national service evaluation of resuscitative thoracotomy practice outside the operating theatre. Resuscitation. 2026;229:111371. https://doi.org/10.1016/j.resuscitation.2026.111371
- Halden Hutchinson-Bazely, “JC: Pre Hospital Resuscitative Thoracotomy for Traumatic Cardiac Arrest – data from 21 years at London’s Air Ambulance,” in St.Emlyn’s, March 6, 2025, https://www.stemlynsblog.org/laa-resuscitative-thoracotomy/.
- Zaf Qasim, “JC: Pre-hospital thoracotomy – what can we learn? St Emlyn’s,” in St.Emlyn’s, February 8, 2022, https://www.stemlynsblog.org/jc-pre-hospital-thoracotomy-what-can-we-learn-st-emlyns/.
- Ter Avest E, Kocierz L, Alvarez C, Hurst T, Ballard D, Lockey DJ, Christian MD, Perkins ZB. Improving decision-making for prehospital Resuscitative Thoracotomy in traumatic cardiac arrest: a data-driven approach. Crit Care. 2025 Nov 13;29(1):485. doi: 10.1186/s13054-025-05705-z. PMID: 41233917; PMCID: PMC12613375.
- Ep 298 – Traumatic Cardiac Arrest: Rethinking Resuscitative Thoracotomy with Laura Kocierz at Trauma 2030 Sep 26, 2026
- Weegenaar, Perkins and Lockey — Traumatic cardiac arrest position statement (2024) Scandinavian Journal of Trauma, Resuscitation and Emergency Medicine. 2024;32:139. https://link.springer.com/article/10.1186/s13049-024-01304-z
- Natalie May, “JC: EAST – Let’s Be Blunt About ED Thoracotomy,” in St.Emlyn’s, August 1, 2015, https://www.stemlynsblog.org/jc-east-lets-be-blunt-about-ed-thoracotomy/.
Cite this article as: Simon Carley, “TETRiS: Resuscitative thoracotomy, traumatic cardiac arrest and the importance of patient selection,” in St.Emlyn’s, November 18, 2026, https://www.stemlynsblog.org/?p=76045.

