Research Roundup – March 2026

Classic medicine: running high cost RCTs that are too small to give real answers

Préterre C, Gaultier A, Obadia M, Vignal C, Mourand I, Plat J, Sablot D, Gaudron M, Rodier G, Godeneche G, Urbanczyk C, Marc G, Massardier E, Adam S, Boulanger M, Marcel S, Mechtouff L, Ronzière T, Calvière L, Godard-Ducceschi S, Barbin L, Lebranchu P, Guillon B; THEIA collaborators. Intravenous alteplase versus oral aspirin for acute central retinal artery occlusion within 4·5 h of severe vision loss (THEIA): a multicentre, double-dummy, patient-blinded and assessor-blinded, randomised, controlled, phase 3 trial. Lancet Neurol. 2025 Nov;24(11):909-919. doi: 10.1016/S1474-4422(25)00308-4. PMID: 41109232

Central retinal artery occlusion has always been frustrating for me. At conferences, people talk about activating stroke teams, immediately involving ophthalmology, and using tPA. In real life, at least in Canada, no one seems interested in doing anything. (Occasionally someone will talk about rubbing an eye, but the chance of that doing anything seems to be lower than Mariana Trench.) This is a well designed, placebo controlled, double-dummy, RCT comparing alteplase to aspirin in 70 patients with confirmed CRAO within 4.5 hours. Unfortunately, it does nothing to lessen my frustration because it is woefully underpowered (they designed the study assuming they would see a 30% absolute improvement in symptoms). Technically, the study is negative, but the percentage of patients with what they considered improved vision at 1 month was 66% with alteplase versus 48% with aspirin. An 18% absolute would clearly be clinically important, and so although this trial is talked about as being negative, I think it leaves us exactly where we started: not having any clue what the best therapy is, and in need of a large properly designed trial. There are a bunch of other details that you could get into with this paper, but I am not sure they really change that overall conclusion. Obviously, it is not easy to perform a large study of a rare condition like CRAO. However, when you design your study with an unrealistic power calculation, you end up doing 90% of the work for 0% of the payoff, and now someone just needs to go through this exact same effort again. At some point, the medical research community is going to have to learn this lesson. 

Bottom line: This trial was statistically negative, but shows a potentially important improvement when using alteplase over aspirin for CRAO. I have no idea how to proceed, and expect the answer will be a crapshoot completely dependent on what specialists are working any given day.

Non-Invasive Management of Blunt Traumatic Pneumothorax

Harrison M. Non-Invasive Management of Blunt Traumatic Pneumothorax-a Meta-Analysis. Emerg Med Australas. 2025 Dec;37(6):e70164. doi: 10.1111/1742-6723.70164. PMID: 41163445

I have spent a lot of time preaching conservative management for spontaneous pneumothorax. Although I often leave them out of my talks, for fear of complicating things, I also frequently take the conservative approach with blunt traumatic pneumothoraces. This is a systematic review and meta-analysis which includes 3 RCTs, 4 prospective observational studies, and 7 retrospective studies, with a total of 1550 patients. The rate of progression of pneumothorax on subsequent imaging was statistically unchanged – 12% in conservative management and 8% of those with a chest tube placed. 11.9% of patients treated conservatively ultimately had a chest tube placed. That number is remarkable for 2 reasons. First, it means that 90% of patients with traumatic pneumothoraces are being safely managed without an invasive procedure. More interestingly, 10.4% of the chest tube group also had to have another chest tube placed. In other words, putting a chest tube in up front did not change the need for a future chest tube at all! Of course, only 3 of these trials are randomized, and so there is almost certainly selection bias – in that the sicker patients are the ones selected for chest tubes – but I think this data pretty convincingly says that if you don’t think a chest tube is needed, there is a very good chance you are correct. Like most medical data, there is an extreme bias presented in the lack of reporting on harms. We know chest tubes have a very high complication rate, and when the harms are not reported in studies, it is impossible to perform a harm/benefit analysis. I will note that multiple studies have used conservative management in patients on positive pressure ventilation, and it is still frequently a successful strategy, but the numbers shift slightly, such that 18% of conservative management patients need a future chest tube placed as compared to only 9% of those with a chest tube placed immediately. 

Bottom line: Conservative management is a good option even for patients with blunt trauma, as long as you are being selective about the patient you chose, and consider factors like length of transfer to a trauma center, positive pressure use, and ease of chest tube placement in whatever setting the patient might be monitored. 

Identifying ventricular fibrillation using ultrasound

Gaspari R, Lindsay R, She T, Acuna J, Balk A, Bartnik J, Baxter J, Clare D, Caplan RJ, DeAngelis J, Filler L, Graham P, Hill M, Hipskind J, Joseph R, Kapoor M, Kummer T, Lewis M, Midgley S, Nalbandian A, Narveas-Guerra O, Nomura J, Sanjeevan I, Scheatzle M, Schnittke N, Secko M, Soucy Z, Stowell JR, Theophanous RG, Tozer J, Yates T, Gleeson T. Incidence and Clinical Relevance of Echocardiographic Visualization of Occult Ventricular Fibrillation: A Multicenter Prospective Study of Patients Presenting to the Emergency Department After Out-of-Hospital Cardiac Arrest. Ann Emerg Med. 2025 Oct;86(4):328-336. doi: 10.1016/j.annemergmed.2025.04.014. Epub 2025 Jun 30. PMID: 40590825

This is a multicenter prospective observational trial that looked at a convenience sample of adult patients with atraumatic out of hospital cardiac arrest and who happened to have a simultaneous bedside echo and ECG during one of the first 3 CPR pauses in the emergency department. They include 811 patients, and their primary outcome is that they found 43 patients (5%) who had ‘occult v fib’ – v fib on echo but not ECG – during 1 of the first 3 rhythm checks. There seems to be something wrong with their dataset, but I can’t figure out exactly what it is. Only 54.5% of patients with ventricular fibrillation on ECG were defibrillated! That is abysmal, and I have a hard time believing that it represents real world practice. I think it suggests a significant bias in their data collection. Another oddity in this data is that apparently 80% of patients with ‘occult v fib’ actually had PEA rather than asystole on their ECG. I don’t know how that is possible, at least without a complete redefinition of what ventricular fibrillation means. We have proof, using the gold standard test, that there is not ventricular fibrillation, and you are calling it ventricular fibrillation on echo anyway? Does that make sense? To me, that sounds a lot like there are just a very large number of false positives on ultrasound. I might also question their definition of defibrillation success, because if I don’t trust the diagnosis of v fib on ultrasound, I can’t really trust the diagnosis of resolution of v fib. That being said, ROSC was actually more common for occult VF than ECG VF (40% vs 25%). Putting aside issues of the accuracy of their definitions, looking at this non-randomized data, I am almost 100% sure that the difference in ROSC is due to confounders. For example, 37% of the patients with apparent ECG VF actually had an initial rhythm of asytole with EMS, as compared to only 5% in the occult VF group. The occult VF group also had 10% more bystander CPR, and they continued to resuscitate the occult VF group 10 minutes longer than the real VF group, even though resuscitations should continue longest in patients with true shockable rhythms. I don’t think you can see all the differences here, but I think there are clearly fundamental differences between these two groups that led to different decisions and different outcomes that are independent of the ultrasound findings.

Not really related to the critical appraisal, but it also really bothers me that there were 52 patients with ECG proven v fib where they still did an echo. You absolutely should not be performing an echo if there is any evidence of ventricular fibrillation on the strip; you should be defibrillating immediately. Even in the cases of transesophageal echo (which was less than 5% of this group), you really should not have time to record a clip if you are performing high quality resuscitation.

What should we do with this data clinically? Personally, I definitely will not be labelling PEA as ventricular fibrillation – that makes no sense to me. However, I have had a number of cases of ventricular fibrillation which looks a lot like asystole by the time you get to your 3rd or 4th shock. If the case started as a shockable rhythm, it makes sense to continue to treat it as a shockable rhythm if there is any doubt, and movement on echo can certainly push me in that direction. If the presenting rhythm was asystole, I don’t think there is significant downside from defibrillation, but I also don’t think you are going to find nearly as many VF cases as they claim here. 

Random question for the echo nerds out there: are there any views of the heart that would be better or worse for picking up fibrillation? Is fibrillation truly random, or does it have a direction or wave component that might be more visible in one direction or another?

Final comment: I do like that the Canadian hospitals were in the “Ottawa province”. I am not sure if that is because our colleagues in Ottawa are trying to take over, or simply a lack of proof reading.

Bottom line: This study will almost certainly be misrepresented by people who just read the abstract. The data is incredibly messy, and should be taken with a massive grain of salt. However, if you believe ventricular fibrillation is a possibility clinically, but the monitor looks like asystole, I think it is reasonable to shock if someone sees fibrillations on an echo. 

The harms of ignoring negative trials

Shi AC, Taylor T, Huang CC, Singhal AB, Goldstein JN, Bevers MB, Hou PC. Early Intensive Blood Pressure Reduction After Intracerebral Hemorrhage Is Associated With Worse Functional Outcome: The Risk of Overshooting Blood Pressure Goals. Ann Emerg Med. 2025 Dec 9:S0196-0644(25)01303-4. doi: 10.1016/j.annemergmed.2025.10.009. Epub ahead of print. PMID: 41369631

In one of the funny quirks of evidence based medicine, experts always suggest aggressive management of blood pressure in the setting of intracerebral hemorrhage despite the existence of 2 large RCTs (ATACH 2 and INTERACT 2) that pretty convincingly show no benefit. Of course, if there is no benefit, it is highly likely that you are going to cause harm. This is a retrospective look at 420 adult patients with spontaneous intracerebral hemorrhage. Based on current guidelines, they defined a target blood pressure as less than 150 systolic. 63% of patients were given antihypertensive therapy and 71% of patients hit the 150 mmHg target. Modified Rankin scores at discharge were worse in the group of patients with ‘target’ blood pressures, as compared to those that remained hypertensive (odds ratio of 2.53 for having a poor outcome). This seems to be the result of blood pressure overshooting. Half of all patients who had blood pressure lowered “overshot”, meaning they had at least one recorded blood pressure less than 120 systolic. Overshooting occurred at equal rates whether or not a bolus of antihypertensive was used. The harms of lowering blood pressure seem to be entirely in the patients who overshot the targets. However, given that there is no proven benefit of managing hypertension in these patients, this is all pure harm. I wouldn’t put too much faith in this retrospective dataset, but it makes sense based on the high quality RCTs that exist. If you want to know whether you should reduce blood pressure in ICH go back and read INTERACT2 and ATACH2. The answer is pretty clear: no. 

Bottom line: This retrospective cohort demonstrates the potential harms of blood pressure management in intracerebral hemorrhage. Neurologic outcomes were worse in the group of patients with ‘guideline concordant’ blood pressure control, likely driven by overshooting to systolic blood pressures less than 120. You should base your practice on the existing large RCTs, which very convincingly show no benefit from blood pressure management. 

I really don’t understand

Kerr G, Chown A, Mercuri M, Clayton N, Mercier É, Morris J, Jeanmonod R, Eagles D, Varner C, Barbic D, Parpia S, Buchanan IM, Ali M, Kagoma YK, Shoamanesh A, Engels P, Sharma S, Worster A, McLeod S, Émond M, Stiell I, Papaioannou A, de Wit K. Applying the Canadian Head CT Criteria to Older Adults Seen in the Emergency Department After a Fall. J Am Geriatr Soc. 2026 Feb;74(2):509-515. doi: 10.1111/jgs.70191. Epub 2025 Oct 31. PMID: 41170758

The author list here is a veritable who’s who of Canadian emergency medicine researchers, but despite having more total brain power than a small country, they have published a paper that makes absolutely no sense to me, and a lot of people seem to be misinterpreting. Despite not applying to the elderly population, they look at the inclusion criteria for the Canadian CT head rule, and conclude that if there is uncertainty about those criteria (ie, if the patient doesn’t know if they lost consciousness), there is a high risk of intracranial hemorrhage, and you should proceed with the CT. (I don’t think the numbers matter, but for reference the risk of ICH was 7.7% if you hit your head and definitely met the CT head rule criteria, 2.5% if you hit your head and didn’t meet the criteria, 7.6% if you hit your head and were unsure about the criteria, 4.6% of you were unsure if you hit your head, and 0.8% if you didn’t hit your head.) However, this paper seems fundamentally flawed to me, in that they looked at the inclusion criteria, but not the exclusion criteria for the Canadian CT head rule. This data includes patients on anticoagulants! What clinician is skipping a CT head in an elderly patient (or any patient) with a known head injury and anticoagulation? They don’t parse the data any further, so for all we know 100% of the bleeds were in the 25% of the population on blood thinners, and the risk is negligible in the patient you would actually consider skipping the CT. No matter what a patient’s age, if there is no indication for immediate neurosurgery, no anticoagulation, and no antiplatelet medication, you are not going to change your management no matter what the CT shows. Those are the patients you should be thinking about forgoing CT. This data, by including a large percentage of patients on anticoagulation, seems completely irrelevant to me.

This study would have been interesting if it looked at a group of patients in whom the CT head rule actually applied. If you have a 55 year old who hit their head, but they are uncertain about loss of consciousness, what is their risk? I would also love to see a study that looks at the actual Canadian CT head rule in an elderly population (just ignoring the age criteria). Ie, what is your risk of hemorrhage if the only reason that you fail the rule is age? But this weird combination focusing only on the inclusion criteria in a population where the rule clearly doesn’t apply makes no sense to me.

Also, we have an RCT that shows that using the Canadian CT head rule results in worse overall medical practice and harm to patients – so maybe just stop using the rule?

Bottom line: I expect you will hear about this paper, and people will try to convince you to CT more (if that is even possible), but this data should not in any way be used to shape your clinical practice, because it does not apply to a patient population in whom you would ever consider forgoing CT. 

Food as a Component of Patient-Centred Care

Barrington V, Carter V, Tagg A, Hitch D. Food as a Component of Patient-Centred Care in Emergency Departments: Preliminary Findings. Emerg Med Australas. 2025 Oct;37(5):e70126. doi: 10.1111/1742-6723.70126. PMID: 40890898

Despite being places designed around healthcare, hospitals are often dismally lacking when it comes to the most basic human health needs such as food, water, and sleep. Hopefully we all now know that patients, even if they might turn out to be surgical, should all be encouraged to eat and drink during their emergency department stays. Unfortunately, even if you buy into that concept in theory, to find palatable food one usually needs to get out of the emergency department (and probably the entire hospital). And that doesn’t even account for culturally appropriate diets or other variety that is important for sick patients to feel comfortable. This is a survey of patients and healthcare providers from Australia that basically states the obvious for anyone who has worked in an emergency department: our food options, if they even exist, suck. I find it so embarrassing and disheartening to work in departments that I would not want to spend a minute in as a patient. Our healthcare spaces are horrendously designed. They lack privacy. They make it impossible to sleep. We deny patients basic necessities like palatable food and drink. We need to stop spending millions of dollars on stupid innovations – like putting CT scanners in ambulances – and divert those funds to ensuring that we create healthcare spaces that we would be proud to show our mothers. We need to actually care for people, not just treat diseases. 

I think we can all empathize with the patient comments:

• “In emergency… food was cold and hard, not presentable… I didn’t even know what they were being served.”
• “It looked like dog food was cold and it had gone hard… just in this one big lump.”
• “Cold, unappetising food”

How many medical diagnoses do you just not know?

Cook K. Rash following a round of golf. Emerg Med J. 2026 Feb 24;43(3):150-166. doi: 10.1136/emermed-2025-215366. PMID: 41735021

I include this not for clinical appraisal, but just a reminder of how much I don’t know, even when it seems like my area of expertise. As a middle-aged white male doctor, golfer’s vasculitis (also known as exercise induced purpura) feels like a diagnosis that should have known about. It presents as a non-blanchable rash, usually around the ankles, following prolonged walking or standing in warm weather. Most importantly, there will not be fever or systemic features. The main reason to know about it is that it is entirely benign, and doesn’t require investigation or treatment (aside from perhaps rest and elevation of the legs).  I am almost certain that I have seen this before, but I definitely did not make the correct diagnosis.

Disposable female urinal bottle

Booth S, Ellis P, Lyttle MD, Lochab S, Pegrum J, Thomas S. Disposable female urinal bottle (the UniWee) improves patient experience for immobile women with lower limb fractures. Emerg Med J. 2025 Apr 22;42(5):326-333. doi: 10.1136/emermed-2024-214181. PMID: 40081970

This is another paper that doesn’t really allow for critical appraisal. It is what Jerry Hoffman and Rick Bukata used to refer to as a “Mikey likes it trial”. It addresses a simple, but often overlooked area of gender imbalance in medical care: the ability to use a urinal. Toileting with a hip or pelvic fracture is a problem, but for men, a urinal is an option that can prevent the need for diapers or a catheter. The provenance of the ‘female urinal’ is unclear, but it is simply a modified standard male urinal, and allows females to urinate while either sitting or lying down. It is made by cutting around the neck of the standard cardboard urinal to create a wider mouth and posterior section that can be slid under the perineum. This was offered to patients with female anatomy in hospitals in South West England, and they present the results of a survey of 103 patients and 118 healthcare professionals. For the most part, it was liked (but not loved). About 70% of patients said that it helped them avoid the pain that is involved with using a bedpan, and is more dignified than using a bedpan, although only 50% thought it increased their privacy or ability to urinate without help. 74% of patients said they would suggest the urinal to other patients, while 12% disagreed. 78% of healthcare professionals (mostly nurses) said they would recommend the use of the urinal, while 13% disagreed. It certainly is not a perfect solution, but I thought it was a very interesting alternative to catheters and bedpans, and it is probably worth discussing with our nurses so that patients are aware of the option. I would love to hear feedback from nurses and female patients, because I feel very unqualified to judge this rather ghetto appearing innovation. Also, if females can urinate in zero gravity environments like the space station, should we not be able to use our existing suction and some kind of long-existing space technology to make this easier on our patients?