The First10EM Research Roundup for September 2025

It has been a while since I have critically appraised. Life without medical journals is sort of nice, but I guess it just isn’t me, so I am back and nerdier than ever. That means there are some strange selections this time around, but there are bound to be at least a few people out there that find these papers as fascinating as I do. 

Antibiotics in strep pharyngitis: another example of why non-inferiority trials are clearly inferior 

Gualtieri R, Verolet C, Mardegan C, Papis S, Loevy N, Asner S, Rohr M, Llor J, Heininger U, Lacroix L, Pittet LF, Posfay-Barbe KM. Amoxicillin vs. placebo to reduce symptoms in children with group A streptococcal pharyngitis: a randomized, multicenter, double-blind, non-inferiority trial. Eur J Pediatr. 2024 Nov;183(11):4773-4782. doi: 10.1007/s00431-024-05705-1. Epub 2024 Aug 31. PMID: 39215861

I have been holding off on talking about this paper for a while, because I am part way through a very in-depth look at all the evidence for antibiotics in strep pharyngitis. However, a few months have passed without me making progress on that project, so I figured we might as well discuss this now. 

I have not used antibiotics for strep throat in Canada since graduating residency. (However, I have used antibiotics in settings with higher rates of rheumatic fever, such as rural New Zealand.) The most recent study on this topic is a double-blind non-inferiority trial that compared 6 days of amoxicillin (50mg/kg divided BID) to placebo in healthy children with a fever, pharyngitis, McIsaac score 3 or more, and positive rapid antigen test for group A strep. (Gualtieri 2024) Unfortunately, the trial ended up under-powered, both because of COVID and because a lot of parents refused to participate. Being under-powered is a huge deal in a non-inferiority trial design, because the non-inferiority conclusion is based entirely on the size of the confidence intervals, so you can’t actually rely on the statistics in a non-inferiority trial that didn’t hit its target enrollment. They ended up with 88 children, and there was no real difference in their primary outcome of fever duration (21.7 hours with amoxicillin and 24.6 hours with placebo.) They can’t technically claim non-inferiority, because the confidence intervals are too big, but they clearly would have reached a conclusion of non-inferiority if they had enrolled as many patients as they originally planned. There were no differences in symptoms between the groups, no differences in paracetamol or NSAID use. Treatment failures included scarlet fever, otitis media, and retropharyngeal abscess, and were not statistically different, but clearly warrant a bigger study, as they occurred in 13% of the placebo group vs 5% of the amoxicillin group (p=0.1). (These numbers are way higher than expected. My guess is that it is a mistake to include otitis media as a complication, because almost any kid that still has a fever at 5 days will find a doctor who will look in their ear and decide it is red just as an excuse to give antibiotics, even though antibiotics are clearly not indicated for otitis media either.) This trial clearly shows no benefit in their primary outcome of fever duration, but do we really care about fever duration? We are left with a big question around suppurative complications, which were double here, but at rates way higher than I have ever seen, and could easily be due to chance alone given that it is a statistically insignificant secondary outcome.

Bottom line: In a setting with essentially no rheumatic fever, this trial seems to fit with prior studies which indicate that the harms of antibiotics probably outweigh any benefit that might exist. Personally, when working in urban Canada, I don’t use antibiotics, which means I also don’t test for strep. (A much more in depth review is coming, and there is likely some nuance, but based on what I have seen so far, the conclusion is probably the same.)

Related: You don’t understand non-inferiority trials (and neither do I)

Maternal and neonatal outcomes following resuscitative hysterotomy

Leech C, Nutbeam T, Chu J, Knight M, Hinshaw K, Appleyard TL, Cowan S, Couper K, Yeung J. Maternal and neonatal outcomes following resuscitative hysterotomy for out of hospital cardiac arrest: A systematic review. Resuscitation. 2025 Feb;207:110479. doi: 10.1016/j.resuscitation.2024.110479. Epub 2024 Dec 29. PMID: 39736393

This paper is not going to change your day to day clinical practice, but might help set expectations if you are faced with performing a perimortem hysterotomy. This is a systematic review, which follows PRESPERO methodology, although I doubt that methodology matters much when your 42 publications contain a total of 66 women and 68 neonates from 1 cohort study, 3 case series, and 38 case reports. Maternal survival was a little lower than I thought, at only 4.5%. Newborn survival was pretty good, at 45%. Although we want to start this procedure as soon as possible, there was survival with good neurologic function in a mother who had the procedure 29 minutes post arrest and a neonate when the procedure was started 47 minutes post arrest. The results actually run counter to the teaching I have heard repetitively over the last decade. There has been a push to rename this a “perimortem hysterotomy” instead of the classic “perimortem c-section”, based on the thought that we are performing the procedure to save the mother’s life, rather than the fetus, but the neonatal survival here was 10 times the maternal survival. (On reflection, I guess that makes sense, given that it is the mother who died of some clearly significant pathology, and the healthy baby is presumably just along for the ride.) That being said, we are talking about 66 cases, and there is going to be massive bias in these numbers. Most cases, even when dramatic procedures are performed, aren’t published, and the cases you choose to write up generally will look very different than the cases you choose not to publish. My guess is that the bias would be in the direction of survival, and so unfortunately we should expect even worse outcomes in real life.

Bottom line: Don’t be too hard on yourself if you do the right thing, perform a resuscitative hysterotomy, and there are no survivors. 

SSRIs as a cancer treatment?

Li B, Elsten-Brown J, Li M, Zhu E, Li Z, Chen Y, Kang E, Ma F, Chiang J, Li YR, Zhu Y, Huang J, Fung A, Scarborough Q, Cadd R, Zhou JJ, Chin AI, Pellegrini M, Yang L. Serotonin transporter inhibits antitumor immunity through regulating the intratumoral serotonin axis. Cell. 2025 Jul 10;188(14):3823-3842.e21. doi: 10.1016/j.cell.2025.04.032. Epub 2025 May 21. PMID: 40403728

I don’t usually include things like this in the Research Roundup, because there are so many ways that new therapies might fail at the clinical trial level, but I thought I would throw in this report, first because it is good to remain optimistic about basic science and the pipeline of novel therapeutics that might be available in the future, and second because it is just plain interesting, and remaining curious about medicine is a good way to prolong one’s career. Basically, based on physiologic pathways and some animal models, SSRIs might be able to reduce tumor growth in certain cancers. (If true, given how common SSRIs are, we probably should be able to find hints of this in observational data sets.) As I said, this will almost certainty turn out not to be true, but even when it is not going to change your day to day practice, I think it is probably helpful for all doctors to review some of the biologic pathways we have not considered since medical school, and marvel at the complexity of human life. 

Deliberate clinical inertia: A paradoxical strategy to improve patient flow

Keijzers G. Deliberate clinical inertia: A paradoxical strategy to improve patient flow? Emerg Med Australas. 2025 Apr;37(2):e70016. doi: 10.1111/1742-6723.70016. PMID: 40045755

Deliberate clinical inertia: the conscious process of ‘doing nothing’ as a positive response to clinician and system intervention bias.

The title of this paper caught my eye instantly. It probably shouldn’t have surprised me that it was written by one of my favourite people in medicine. (Although I am a little concerned that by including his paper I am forcing myself to pronounce his name properly on the podcast.) This is a short essay discussing the lost art of doing nothing in medicine. In response to the ever-increasing practice in emergency medicine of having tests sent before the patient is even assessed by a physician, he proposes a brilliant study: “randomise patients to ‘assessment by senior clinician before any tests can be ordered’ versus ‘usual care’.” It might be difficult to achieve, and unfortunately we have trained modern patients to expect tests. I love practicing minimalist medicine, but find myself feeling more and more pressured by patients to order tests I don’t think are necessary. He highlights that “deliberate clinical inertia is not merely an exercise in restraint or avoiding testing and interventions; it is an opportunity to enhance patient engagement and satisfaction.” I think this is entirely true. When I see bouncebacks from other emergency departments, I frequently hear patients say “all they did was send a bunch of tests and tell me I was fine”. Although the workup performed was probably adequate, if not superb, the patient did not feel heard, and was therefore unsatisfied with their care. Conversely, patients who feel like they have their doctor’s undivided attention often don’t care at all about the specifics of the workup. 

He includes a list of practices with good evidence for conservative management that might change practice for some people:

• Buddy tape rather than cast for Boxer’s fractures
• Conservative management of pneumothorax
• Reducing IV placement in the ED
• Reduction in coagulation studies which essentially never change management
• Managing alcohol intoxication with observation alone (why are you ordering bloodwork or an IV, when you could just wait for the patient to sober?)
• Oral antibiotics over IV for cellulitis

Clearly, Dr. Keijzers owns the same soapbox I do.

Bottom line: I definitely buy this argument. The primary efficiency tip I always give to trainees is to avoid unnecessary tests. Even when they seem minor, adding extra patient touches will grind a shift to a halt. 

The floppy iris syndrome

Zaman F, Bach C, Junaid I, Papatsoris AG, Pati J, Masood J, Buchholz N. The floppy iris syndrome – what urologists and ophthalmologists need to know. Curr Urol. 2012 May;6(1):1-7. doi: 10.1159/000338861. Epub 2012 Apr 30. PMID: 24917702

There is absolutely no reason that anyone in the emergency medicine or critical care worlds need to know this, but for some reason it tickled my fancy, and learning interesting facts about the human body just can’t be bad for a doctor. Intraoperative floppy iris syndrome is apparently something that happens in patients who are taking tamsulosin. The alpha blockade results in incomplete pupil dilatation, which can impair removal of the cataract and insertion of the new lens. Can I stretch this to try to make the information clinically relevant? Well, I am not sure you should ever prescribe tamsulosin for nephrolithiasis, but if that is your practice, adding a question about upcoming cataract surgery might help your patient? Realistically, this information is useless to us. I just like the name “floppy iris syndrome” and I hope to officially write it on a patient’s chart somehow (although seeing as this is an intraoperative diagnosis, I am not exactly sure how I am going to make that happen.)

Bottom line: This might be the time for you to unsubscribe from these emails. 

Our endless quest to treat the number not the patient

Endo A, Yamakawa K, Tagami T, et al. Efficacy of targeting high mean arterial pressure for older patients with septic shock (OPTPRESS): a multicentre, pragmatic, open-label, randomised controlled trial. Intensive Care Med. 2025 May;51(5):883-892. doi: 10.1007/s00134-025-07910-4. Epub 2025 May 13. PMID: 40358717

If elderly patients are more likely to have baseline hypertension, or untreated/ undiagnosed hypertension, does it make sense to target higher MAPs in sepsis? This open-label RCT from Japan compared a target MAP of 80-85 to the standard target of 65-70 in ICU patients aged 65 and older with sepsis. It was stopped early because of harm detected at the interim analysis, with 516 patients enrolled out of the 752 target. 90 day mortality was 39% in the high-MAP group as compared to 29% in the control group (ARR 10.7%, 95% CI 2.6-18.9%). Although none were statistically significant, all adverse events (arrhythmias, ischemia, and bleeding) were more frequent in the high-MAP group. I tend to believe that targeting a higher MAP would result in more adverse events, because basically every time we try to tightly control numbers rather than treating patients we end up causing harm. That being said, given the usual outcomes in intensive care research, there is just no way that a small change in blood pressure target is going to kill one in ten patients. This is either a purely statistical deviation, made worse by the fact that an already too small trial was stopped early, or the effect of some kind of unseen bias in an unblinded trial. Either way, given that there is clearly no hint of benefit, there is no reason at all to change our current practice away from the target MAP of 65. (Conversely, don’t worry too much if you overshoot, because both of these groups had a mean MAPs very close to 80 throughout most of the trial, which is another reason not to believe the 10% increase in mortality.)

Bottom line: Although I don’t think the massive mortality difference is believable, everything in this trial points at harm from targeting higher MAPs. 65 is the target. Focus on clinical signs of good perfusion rather than chasing numbers.

Full OPTPRESS write up here

Virtual reality and music therapy? Does that sound like any emergency department you have ever worked in?

Dönmez S, Şener A, Işık Nİ, Koca S, Akbaş İ. Analgesic and Anxiolytic Effects of Virtual Reality and Classical Music Therapy on Tension-Type Headache: A Randomized Controlled Trial. Ann Emerg Med. 2025 Jun 4:S0196-0644(25)00278-1. doi: 10.1016/j.annemergmed.2025.04.032. Epub ahead of print. PMID: 40464717

This single-centered randomized trial from Turkey compared dexketoprofentrometamol (and NSAID) 25 mg IV alone to 25 mg of dexketoprofentrometamol plus a VR experience with classical music in 140 ED patients with tension-type headaches. The VR experience was YouTube based nature scenery (which I appreciated was from Canada, although I found the quality of the images to be quite poor) accompanied by classical music. The VR group had significantly greater reductions in pain at 30, 60, and 120 minutes. The absolute differences were pretty big, and bigger than is considered clinically significant (eg., at 120 minutes the control group had moved from a baseline of 80/100 to 46, whereas the VR group had moved from 80 to 12). Patient mood was also significantly better in the VR group. Only 3% of the VR group needed rescue meds, compared to 46% in the control group. This is an interesting study, and the results are believable (distraction is good for all pain, and relaxation is almost certainly helpful in tension type headaches). Of course, this is a single center unblinded trial with a subjective outcome and results that sound almost too good to be true. You can almost guarantee some degree of bias in the results. There are no harms reported here, but they did limit use of this intervention. I could imagine a migraine could actually be made a lot worse by a VR headset and music. What this study really makes me consider is the current treatment of these patients. The emergency department waiting room is basically the exact opposite of beautiful nature scenery and serene music. As compared to a simple baseline of a quiet, private room, might our current treatment be making patients worse? Does the fancy VR rig actually add anything, or would we see the same results with a quiet room with relaxing music? Or perhaps the best plan is to just discharge patients home to their own comfortable environments?

Bottom line: A single center study with extremely high risk of bias clearly isn’t practice changing, but it doesn’t seem like you need a practice changing paper to consider and try to improve the environment in which you care for patients. 

Does ozempic make you puke?

Chen YH, Zink T, Chen YW, Nin DZ, Talmo CT, Hollenbeck BL, Grant AR, Niu R, Chang DC, Smith EL. Postoperative Aspiration Pneumonia Among Adults Using GLP-1 Receptor Agonists. JAMA Netw Open. 2025 Mar 3;8(3):e250081. doi: 10.1001/jamanetworkopen.2025.0081. PMID: 40036031

I have spent a lot of time recently talking about the silly NPO or fasting rules that most emergency departments enforce on their patients, at great cost and with no benefit. One of the fears raised by anesthesiologists is that even if shorter fasting did not increase aspiration risk in the past, the rise of GLP-1 agonists and their associated delayed gastric emptying means that we are going to see far more aspiration in the future. Some have told me that fasting rules are going to have to become even more stringent. This is the first data I have seen on the topic, a retrospective cohort based on an American private insurance database. They looked at adult patients undergoing one of 14 common surgical procedures, and compared the 30 day post-operative rate of aspiration pneumonia among those patients on any type of GLP1 agonist (dulaglutide, exenatide, liraglutide, or semaglutide) to controls without a prescription in the 30 days before surgery. A total of 366 476 patients were included in the study. There were no statistical differences in the rates of aspiration pneumonia. If anything, the rate of aspiration pneumonia seems to be lower in the GLP1 agonist group (OR 0.78 95% CI 0.57-1.06.) Of course, confounders are always an issue in database studies. Patients prescribed GLP1 agonists are very different at baseline than patients who are not prescribed these medications. Although there is a lot of complexity, based on the differences between the groups, I would expect the GLP1 group to be at higher risk of aspiration at baseline, with more diabetes, obesity, older age, and higher percentage of bariatric surgery. This data would be harder to interpret if the risk was higher with GLP1s, but just not significantly so. However, given that the GLP1 group had an overall lower rate of aspiration pneumonia, I am comfortable ignoring concerns about delayed gastric emptying until we see real clinical data suggesting harm, not just anecdotes. 

Bottom line: Although there are anecdotal reports coming from anesthesia about an increased risk of aspiration with GLP1 agonists, that concern does not appear to be supported by this large observational data set. 

Comparing apples and oranges: a randomised prospective study

Barone JE. Comparing apples and oranges: a randomised prospective study. BMJ. 2000 Dec 23-30;321(7276):1569-70. doi: 10.1136/bmj.321.7276.1569. PMID: 11124178

I don’t know how it is possible that I never read this paper before. I guess I have the excuse of still being in high school when it was published. Basically, this paper definitely shows that all those science eggheads who tell you that you can’t compare apples and oranges are just dead wrong. The paper starts, “For many years the comparison of apples and oranges was thought to be impossible. Many authors use the analogy of the putative inability to compare apples and oranges as a means of scornfully reviewing the work of others… Our laboratory has been interested in this problem for many years. We attempted numerous pilot studies (unpublished data) but had not accomplished a true comparison until now. At last, successful comparison of apples and oranges has been achieved and is the subject of this report.” They compared 6 red delicious apples to 6 navel oranges, using multiple measurement tools including a tape measure, a scale, and a “Licker scale” from 1 (kind of sweet) to 4 (really very sweet). The fruits are remarkably similar. In fact, the only statistical difference they were able to find was in “involvement of Johnny Appleseed”. They conclude, “this article, certain to become the classic in the field, clearly demonstrates that apples and oranges are not only comparable; indeed they are quite similar. The admonition “Let’s not compare apples with oranges” should be replaced immediately with a more appropriate expression such as “Let’s not compare walnuts with elephants” or “Let’s not compare tumour necrosis factor with linguini.””

Cheesy Joke of the Month

If you rearrange the letter of MAILMEN…

They get really upset