The December 2024 Research Roundup

Another month, another set of articles. Perhaps some are useful. Perhaps some will guide your practice. At least one should completely blow your mind.

More research was needed

Taccone FS, Rynkowski Bittencourt C, Møller K, Lormans P, Quintana-Díaz M, Caricato A, Cardoso Ferreira MA, Badenes R, Kurtz P, Søndergaard CB, Colpaert K, Petterson L, Quintard H, Cinotti R, Gouvêa Bogossian E, Righy C, Silva S, Roman-Pognuz E, Vandewaeter C, Lemke D, Huet O, Mahmoodpoor A, Blandino Ortiz A, van der Jagt M, Chabanne R, Videtta W, Bouzat P, Vincent JL; TRAIN Study Group. Restrictive vs Liberal Transfusion Strategy in Patients With Acute Brain Injury: The TRAIN Randomized Clinical Trial. JAMA. 2024 Oct 9. doi: 10.1001/jama.2024.20424. PMID: 39382241

We did liberal versus conservative blood transfusion in brain injury in October, with the HEMOTION trial, but were left in the awkward spot where the results look better with liberal transfusion, but the trial was statistically negative. As a follow up, we have the TRAIN trial, which is another open-label multicenter RCT comparing liberal to conservative transfusion strategies in brain injury (this time not just trauma, but also subarachnoid hemorrhage and intracerebral hemorrhage). The trial includes 850 patients, 60% of which is trauma, and about 20% each SAH and ICH. For their primary outcome, 63% of the liberal group and 73% of the conservative group had unfavorable neurologic outcomes (ARR 10%, 95% CI 4-17%). Mortality was about the same (20.7% vs 22.5%). There were fewer cerebral ischemic events with liberal transfusion, but the other secondary outcomes were essentially unchanged. The big problem with this trial is that it is unblinded and the outcome is subjective. The second biggest issue is that they changed their sample size calculation, and the way they changed it is a little suspicious. Those are big enough red flags that if this were a big practice change, I would suggest holding off. However, all these results are saying is that we should consider keeping hemoglobin above 90 g/L in brain injured patients, and that seems reasonable, especially given that these results are basically saying the same thing as HEMOTION. However, I would not be surprised if future higher quality data proves us wrong.

Bottom line: For now, it looks like there might be a small neurologic benefit from transfusing patients with brain injury and a hemoglobin less than 90 g/L (that’s  9 g/dL for Americans).

Full post here.

Skepticism: a lost clinical art.

DiNubile MJ. Skepticism: a lost clinical art. Clin Infect Dis. 2000 Aug;31(2):513-8. doi: 10.1086/313945. Epub 2000 Sep 14. PMID: 10987714

I present this essay for quiet contemplation by the fireplace with a glass of something delicious in your hand. He starts, “Skepticism is generally regarded in scientific circles as a necessary element of healthy agnosticism, less adversarial than cynicism but much more demanding than faith. Under the weight of solid evidence, the skeptic can be moved toward acceptance and even evolve, albeit reluctantly, into a true believer. However, in our modern era of unprecedented scientific growth, where significant biomedical advancements are daily occurrences, it appears that contemporary physicians have become more willing to accept the “latest and greatest” without careful scrutiny.” He proceeds to discuss, through the lens of an infectious disease specialist, a number of issues with modern medicine. I don’t agree with everything said. For example, I think he is somewhat mistaken in his take on the impact of the placebo effect. However, this is a great essay if you just want to take some time to reflect on the philosophical underpinnings of your medical practice. As an example, I will provide one table of aphorisms he provides:

IV or IO? (Or nothing at all?)

Couper K, Ji C, Deakin CD, Fothergill RT, Nolan JP, Long JB, Mason JM, Michelet F, Norman C, Nwankwo H, Quinn T, Slowther AM, Smyth MA, Starr KR, Walker A, Wood S, Bell S, Bradley G, Brown M, Brown S, Burrow E, Charlton K, Claxton Dip A, Dra’gon V, Evans C, Falloon J, Foster T, Kearney J, Lang N, Limmer M, Mellett-Smith A, Miller J, Mills C, Osborne R, Rees N, Spaight RES, Squires GL, Tibbetts B, Waddington M, Whitley GA, Wiles JV, Williams J, Wiltshire S, Wright A, Lall R, Perkins GD; PARAMEDIC-3 Collaborators. A Randomized Trial of Drug Route in Out-of-Hospital Cardiac Arrest. N Engl J Med. 2024 Oct 31:10.1056/NEJMoa2407780. doi: 10.1056/NEJMoa2407780. PMID: 39480216

Vallentin MF, Granfeldt A, Klitgaard TL, Mikkelsen S, Folke F, Christensen HC, Povlsen AL, Petersen AH, Winther S, Frilund LW, Meilandt C, Holmberg MJ, Winther KB, Bach A, Dissing TH, Terkelsen CJ, Christensen S, Kirkegaard Rasmussen L, Mortensen LR, Loldrup ML, Elkmann T, Nielsen AG, Runge C, Klæstrup E, Holm JH, Bak M, Nielsen LR, Pedersen M, Kjærgaard-Andersen G, Hansen PM, Brøchner AC, Christensen EF, Nielsen FM, Nissen CG, Bjørn JW, Burholt P, Obling LER, Holle SLD, Russell L, Alstrøm H, Hestad S, Fogtmann TH, Buciek JUH, Jakobsen K, Krag M, Sandgaard M, Sindberg B, Andersen LW. Intraosseous or Intravenous Vascular Access for Out-of-Hospital Cardiac Arrest. N Engl J Med. 2024 Oct 31. doi: 10.1056/NEJMoa2407616. PMID: 39480221

We have 2 studies looking at the same question published in the same edition of the NEJM, so I will tackle them together. (You can find more details in the full blog post.) Which is a better first option for vascular access in out of hospital cardiac arrest: IV or IO? (Of course, a more fundamental question might be, if drugs don’t help dead people, does vascular access matter at all?)

PARAMEDIC 3 randomized 6,000 patients (but they were supposed to get to 15,000) with out of hospital cardiac arrest from multiple EMS agencies in the UK to either an IO or IV to start. Their primary outcome was survival at 30 days, and there was no difference between the two groups (4.5% of the IO group and 5.1% of the IV group). Favourable neurologic outcomes were seen in 2.7% versus 2.8%. ROSC and sustained ROSC to hospital handover were both slightly higher with IV (2% absolute). There will be some conversation about the slight difference in ROSC here, but if people are not surviving that conversation is irrelevant (and if anything, this ROSC seems to represent harm). 

IVIO is a similar trial from Denmark, in which 1749 adult patients with out of hospital cardiac arrest were randomized to an IO or an IV as their first method of vascular access. The primary outcome here was ROSC, and there was no difference (30% with IO vs 29% with IV, RR 1.06, 95% CI 0.9-1.24). There were no differences in survival (12% with IO vs 10% with IV) or neurologically intact survival (9% vs 8%).

There is one aspect of both of these trials that I don’t understand. First attempt success was far more successful with IO than IV, but despite that, the time it took for patients to receive medications was identical. That doesn’t make any sense at all to me. If you have an IO placed immediately in 95% of patients, but only get an IV in 65%, how are you not getting medications into the IO group faster? It sort of negates the entire theoretical basis for this trial. The only reason we would think IO might be better is because it is quicker and easier. If the timing is identical, obviously there is not going to be a difference. 

Although there are some problems with these trials, I believe the results. We have lots of evidence that medications are pointless in dead patients. Therefore, the route by which you provide the medication is going to be doubly irrelevant. There was a secondary analysis in the IVIO trial that I think demonstrates this beautifully. They compared clinical outcomes between humeral and tibial IOs, and there was no difference. However, a number of these patients had CTs later to confirm IO placement, and although 100% of tibial IOs were in place, only 70% of humeral IOs were. In other words, even when you don’t successfully get vascular access (and are completely unaware of it), it has no impact on outcomes. It is quite possible that you could have added a ‘no vascular access’ group to both of these trials and seen identical outcomes.

Bottom line: The strict bottom line is that we have 2 trials pretty clearly demonstrating equivalence between IO and IV as the first choice vascular access in cardiac arrest. The larger bottom line is that the route of vascular access is irrelevant if your medications are providing no benefit. 

Shorter courses of antibiotics

Ivankovic M, Schwartz KL. Shorter courses of antibiotics. CMAJ. 2024 Nov 3;196(37):E1266. doi: 10.1503/cmaj.240246. PMID: 39496357

An excellent short article by friend and colleague Maria Ivankovic telling us that shorter is usually better when it comes to antibiotics. The key points made are:

• Antimicrobial resistance (AMR) is an urgent public health threat
• Longer antibiotic courses do not prevent AMR; they increase the likelihood of adverse effects and AMR
• Most acute community-acquired infections can be treated with 5–7 days of antibiotics
• The rule that “shorter is better” has some exceptions
• The shortest effective duration of antibiotic therapy is best for patients and society

I might quibble with the use of otitis media as an example of benefit from longer courses of antibiotics, as 0 days of antibiotics is almost certainly the best course, but the main points are spot on. (Unfortunately, in the same month that this article was published, I saw multiple social media posts from the WHO chastising patients for not taking their full course of antibiotics even if they were clinically better. The WHO has never been good at science, so this doesn’t surprise me, but it does disappoint me.)

Age adjusted D-dimer in high risk patients

Bannelier H, Kapfer T, Roussel M, Freund Y, Alame K, Catoire P, Vromant A. Failure rate of D-dimer testing in patients with high clinical probability of pulmonary embolism: Ancillary analysis of three European studies. Acad Emerg Med. 2024 Nov 1. doi: 10.1111/acem.15032. PMID: 39487597

This paper caught my eye after a recent debate about when it is appropriate to use an age adjusted D-dimer. This is a retrospective look at 3 European PE trials (PROPER, MODIGLIANI, and TRYSPEED), to determine the failure rate of age adjusted DDimer specifically in high risk PE patients. High risk was defined as a Wells score greater than 6 or a revised Geneva score greater than 10. Of the 12,300 patients in the original PE studies, there were 651 patients who were classified as high risk and had appropriate follow-up data. Because these are European studies, these were truly high risk patients, with 31.3% of patients being diagnosed with PE. 70 patients had D-dimer levels below the age adjusted cut-off, and none of these patients had a PE (failure rate of 0% but with a 95% CI 0-6.5%). They included PEs diagnosed up to 3 months after the initial visit as misses, and therefore the study was designed to overestimate the patients’ actual risk at the time of emergency department assessment. Obviously, the numbers are too small to be definitive (the confidence interval extends beyond the test threshold), and this is a retrospective look at data not designed for this purpose, so conclusions should be very limited. I have always been taught that D-dimer is not accurate enough (even at the set 500 mg/mL threshold) to rule out PE in high risk patients. I have not been able to find convincing data to settle the question (please send my way if you have it), but the negative likelihood ratio of a negative D-dimer seems to be about 0.15, so basic math tells us that if your pretest probability is between 30 and 40%, a negative D-dimer only gets you to between 5 and 10% risk, which seems too high. However, there has been some debate about the comparative accuracy of age adjusted and set cut-off D-dimers, and I think this is reassuring data that the age adjusted D-dimer likely performs just as well as a set cutoff even in high risk patients. Personally, I use an age-adjusted D-dimer any time I think that D-dimer testing is appropriate, but contrary to the results seen here, I still think the highest risk patients are too high risk for D-dimer alone, and so I proceed directly to CT. That being said, you can’t get better than a 0% miss rate, so if there is better data out there, or I am being overly conservative in my estimate of a negative likelihood ratio of 0.15, please send data my way, because it would be great if a negative D-dimer got even high risk patients below the test threshold.

Bottom line: This is retrospective data that shouldn’t change practice, but in this population the miss rate with an age adjusted D-dimer in high risk patients was 0%. 

It’s Not Cyclic Vomiting Syndrome Until Dietl’s is Ruled Out

Thom C, Larsen M, Kongkatong M, Ottenhoff J, Moak J. It’s Not Cyclic Vomiting Syndrome Until Dietl’s is Ruled Out: A Case for Point of Care Renal Ultrasound. J Emerg Med. 2024 Oct;67(4):e346-e350. doi: 10.1016/j.jemermed.2024.05.003. Epub 2024 May 16. PMID: 39183114

I have been wanting to sneak a paper on Dietl’s syndrome for many years, ever since a urologist casually dropped the eponym on me, and I had to pretend to know what was going on. (Actually, I am quite comfortable in my ignorance, and so just asked what the hell they were talking about.) This paper is really just 2 case reports of pediatric patients who had been given the provisional diagnosis of cyclic vomiting without any prior imaging of the abdomen, and in whom the emergency physicians found hydronephrosis on bedside ultrasound, with the ultimate diagnosis of Dietl’s crisis. I have actually only seen this in adults, but apparently it is more common in pediatrics. You get chronic partial ureteropelvic junction obstruction, and then something (often over-hydration) leads to an acute increase in hydronephrosis, causing nonspecific abdominal pain, nausea, and vomiting. Aside from imaging, all other tests are likely to be negative. The key is that symptoms will be recurrent. Management will require a surgical procedure, and so diagnosis and referral to urology can save these patients a lot of distress. At the end of the day, I think this paper just highlights how different medical practice is in Europe as compared to North America. No one is diagnosed with cyclic vomiting in North America without imaging, and usually many different imaging studies over many visits. We don’t need to be reminded to look for hydronephrosis, because even if it wasn’t on our differential diagnosis, someone would have caught it incidentally on either a CT or ultrasound at some point. Of course, that is easier to say in an adult population. Chronic abdominal pain is common and usually benign in pediatrics, and we definitely should not be using CT very often, but if a child has more than a couple emergency department visits, an ultrasound is definitely warranted. 

Bottom line: If you have never heard of Dietl’s syndrome, this paper might be a worthwhile read. Otherwise, you will likely catch this incidentally on an ultrasound you were already performing or ordering. 

It really doesn’t matter what you do with most lacerations

Barton MS, Chaumet MSG, Hayes J, Hennessy C, Lindsell C, Wormer BA, Kassis SA, Ciener D, Hanson H. A Randomized Controlled Comparison of Guardian-Perceived Cosmetic Outcome of Simple Lacerations Repaired With Either Dermabond, Steri-Strips, or Absorbable Sutures. Pediatr Emerg Care. 2024 Aug 15. doi: 10.1097/PEC.0000000000003244. Epub ahead of print. PMID: 39141836

I have a long series of articles looking at the science behind laceration repair, and I think the simplest answer is: nothing you do matters. Or, if you don’t like that level of nihilism, you could phrase it as, the human body has amazing mechanisms to repair the skin, and our job is mostly to set the natural healing up for success (aka get out of the way). This is an RCT from a single pediatric emergency department, randomizing children with small linear lacerations (less than 5 cm long, less than 5 mm gap, and less than 12 hours old) to dermabond, steri-strips, or absorbable sutures. The primary outcome was cosmetic appearance as rated by the child’s parent at 3 months. They include 55 patients, and three groups had statistically similar outcomes (although the dermabond group was rated 15 points higher on the visual analog scale, which might be clinically significant, and so this tiny study is just too tiny.) Likewise, although none of the secondary outcomes were statistically significant, the point estimates look worse for sutures in length of stay, pain, and overall satisfaction. At the end of the day, this trial doesn’t add much, because they only enrolled small linear lacerations, and we already knew that these healed no matter what you do. (Honestly, most of the lacerations in this study with a median length of 1.5 cm would have probably had the same outcome with a bandaid). At this point, I have almost entirely abandoned sutures in my practice. Dermabond and/or steristrips will close more than 95% of the lacerations we see.

Bottom line: This tiny single center trial doesn’t add a lot, but gives me another opportunity to pitch abandoning sutures to you. You will be more efficient, cause less pain, have happier patients, and your outcomes will be identical. 

PECARN does good work

Schnadower D, Tarr PI, Casper TC, Gorelick MH, Dean JM, O’Connell KJ, Mahajan P, Levine AC, Bhatt SR, Roskind CG, Powell EC, Rogers AJ, Vance C, Sapien RE, Olsen CS, Metheney M, Dickey VP, Hall-Moore C, Freedman SB. Lactobacillus rhamnosus GG versus Placebo for Acute Gastroenteritis in Children. N Engl J Med. 2018 Nov 22;379(21):2002-2014. doi: 10.1056/NEJMoa1802598. PMID: 30462938

PECARN produces enough research that I could probably cover them every month. After last month’s rant about the probably harmful c-spine rule, I wanted to take a break from decision rules (although there are a few other rules from the PECARN group that require critique), and instead look at a previous RCT from PECARN exploring whether probiotics might improve symptoms in gastroenteritis. This is a multicenter double blind RCT that randomized 971 children aged 3 months to 4 years with acute gastroenteritis (define as 3 or more watery stools per day starting less than 7 days ago) to either L. rhamnosus GG 1×10^10 colony forming units twice daily for 5 days, or matching placebo. The severity of gastroenteritis was measured using the modified Vesikari scale (see chart below), and bad outcomes were defined as any patient with a score of 9 or greater, as scored by the parents, at any time during the 14 days following randomization. There was no difference in the primary outcome (11.8% vs 12.6% of children had moderate to severe gastro, p=0.83), and there were also no differences in any of the secondary outcomes. Although there are all sorts of reasons to think that the microbiome is important and has health impacts, probiotics as they are currently envisioned have tremendous hurdles to overcome. There are millions of bacteria one could choose from, and our mechanism of delivery is ineffective, and just giving a probiotic does not mean it will survive or alter the microbiome at all. Until we develop a much better understanding of the microbiome at a very basic science level, I would expect all of these trials to fail, and for probiotics to remain on the ‘alternative medicine’ shelves in the pharmacies. However, unlike some ‘alternative medicines’, such as homeopathy, which have 0% chance of ever working, probiotics may still some day find their home within medicine. (If you want a fascinating and not at all woo based book about the microbiome, I Contain Multitudes by Ed Yong was a really good read.)

Bottom line: What do you call an alternative medicine that actually works? Medicine. Unfortunately, probiotics are still a long way from earning that title. 

Looks like I was wrong

Jaballah R, Toumia M, Youssef R, Ali KBH, Bakir A, Sassi S, Yaakoubi H, Kouraichi C, Dhaoui R, Sekma A, Zorgati A, Beltaief K, Mezgar Z, Khrouf M, Bouida W, Grissa MH, Saad J, Boubaker H, Boukef R, Msolli MA, Nouira S. Piroxicam and paracetamol in the prevention of early recurrent pain and emergency department readmission after renal colic: Randomized placebo-controlled trial. Acad Emerg Med. 2024 Aug 19. doi: 10.1111/acem.14996. Epub ahead of print. PMID: 39161087

We have tons of data that tells us that NSAIDs are the first line analgesic for acute renal colic. Therefore, it makes sense that NSAIDs would be a key component of discharge analgesia as well. Over the years that I have been practicing, I have noted a trend in the patients who bounced back to the emergency department with uncontrolled pain from their kidney stones: they were all prescribed opioids, but never given NSAIDs. I therefore routinely provide naproxen to patients I discharge home, and because, in my mind, I expect these patients to have pain for the next few days at least, I have been prescribing naproxen as a regularly scheduled rather than as needed medication. It looks like I might be wrong. (Not about the value of NSAIDs, but at least about the value of using them regularly to prevent pain.) This is an RCT from 4 emergency department in Tunisia, in which adult patients with clinically diagnosed renal colic were randomized to piroxicam, paracetamol (acetaminophen), or placebo, to be taken regularly for 5 days in an effort to prevent or control expected pain. They randomized 1383 patients, and there was no difference in the rate of recurrence of renal colic pain (based on phone calls at 7 days), with about 30% of each group reporting a recurrence of pain. Now, not all pain is created equally, and for some reason they didn’t ask how bad the pain was. It is possible that the NSAID group was having 5/10 pain while the placebo group was having 10/10 pain, and we just wouldn’t know. However, there doesn’t seem to be a dramatic difference, as the rate of ED return visits was approximately the same in each group: 20.8% with piroxicam, 23.8% with paracetamol, and 22.9% with placebo. As compared to my experience, that is a very high rate of ED bounce back, and it speaks to the fact that this population does not seem to perfectly match the population of patients I have been sending home with a diagnosis of renal colic. I am a strong believer that imaging is not required in most cases of clear renal colic, clinically speaking, but a lack of imaging could be a weakness in a research setting. Half of these patients were managed without any imaging at all, and of those with imaging, only 40% had a documented stone (the rest had hydronephrosis). We know NSAIDs are incredibly effective when patients have proven stones, but if a large proportion of these patients did not have a stone (see Dietl’s syndrome above), that might dilute the efficacy of the NSAID. The other major limitation of this study is that they don’t have any documentation about outside medication use (although patients were instructed to avoid over the counter medications). Perhaps the pain levels were identical, but the placebo group ended up using a lot more rescue mediation.

Bottom line: This is an imperfect trial, but it is a great effort, and as far as I know is the only RCT to look at prophylactic analgesia post ED discharge. I still think NSAIDs are an essential part of the discharge analgesia plan, but this might change my practice from prescribing them to be taken regularly to instead giving them PRN. 

Are ants better doctors than we are?

Frank ET, Buffat D, Liberti J, Aibekova L, Economo EP, Keller L. Wound-dependent leg amputations to combat infections in an ant society. Curr Biol. 2024 Jul 22;34(14):3273-3278.e3. doi: 10.1016/j.cub.2024.06.021. Epub 2024 Jul 2. PMID: 38959879

This might be the best paper I have ever discussed on the blog/podcast, shared by one of the few people who can easily claim to be nerdier than me, Dr. Aaron Orkin. My mind is honestly blown. Ants practice medicine on each other; advanced medicine that includes highly successful leg amputations. Yes, I know that this paper is entirely irrelevant to your day to day practice of medicine, but I really think that everyone should read this. I learned so much. I have even more questions. They describe a series of experiments, demonstrating that ants can discriminate between types of injury, and in injuries where amputation will improve survival, nest mates will perform an amputation. The ants apparently always followed the same procedure. They “would begin licking the wound before moving up the injured limb with their mouthparts until they reached the trochanter. The nestmates then proceeded to repeatedly bite the injured leg until it was cut off.”

Injuries to the tibia were not helped by amputation (even when performed by the researchers), and ants apparently know this, never amputating, and instead spending a lot more time in wound care. (They essentially lick each other.) Impressively, all 21 ants who had their femur chewed off survived, In comparison, ants that were infected but isolated from other ants (and therefore denied ant to ant therapy including amputation) had a 95% mortality rate. In other words, chewing off the leg is one of the most successful therapies I have ever heard of in all of medicine, and certainly exceeds the benefit of essentially everything we have in human medicine. (The image of orthopedic surgeons trying to learn from this paper, and discussing the best way to improve their masseter strength, so that they too can chew through their patients’ femurs, just tickles me.)

P.S., Who out there knew that there are micro-CT scanners capable of performing CT scans on the microvasculature of ants??!! 

Cheesy Joke of the Month

Two hunters are out in the woods, and one of them collapses.

His friend calls 911. “My friend died! We are in the middle of the woods! What should I do?!”

The operator replies, “Sir, calm down. First, let’s make sure he is really dead.”

There is a silence, and then a loud bang.

The guy comes back on the phone and says, “OK, now what?”