Morgenstern, J. Research Roundup (August 2022), First10EM, August 1, 2022. Available at:
Vitamin C in sepsis: Don’t LOVIT
Lamontagne F, Masse MH, Menard J, et al; LOVIT Investigators and the Canadian Critical Care Trials Group. Intravenous Vitamin C in Adults with Sepsis in the Intensive Care Unit. N Engl J Med. 2022 Jun 23;386(25):2387-2398. doi: 10.1056/NEJMoa2200644. PMID: 35704292
This is a big trial, but the results were so predictable I figured we only needed a brief summary rather than an entire post. I have already covered the ‘metabolic cocktail’ for sepsis at length, with the conclusion that “there is no reason to believe that a metabolic cocktail of vitamin C, thiamine, and steroids improve outcomes in sepsis”. Multiple subsequent trials have confirmed that conclusion, and the LOVIT trial is yet another (but with the distinction of actually demonstrating harm.) This is an international, multi-centre, randomized, placebo controlled trial comparing vitamin C to placebo in 863 adult patients admitted to the ICU with sepsis and requiring a vasopressor. I don’t love that their primary outcome was a composite of death and persistent organ failure (which are qualitatively very different). This primary outcome demonstrated HARM from the vitamin C, with 45% of the vitamin C group dying or having persistent organ dysfunction, as compared to 39% of the placebo group (p=0.01, RR 1.21, 95% CI 1.04-1.40). Death was not statistically significant, although the difference looks similar (35.4% vs 31.6%, RR 1.17; 95% CI 0.98-1.40). This trial could have been so much stronger if it had just been powered for mortality and made a little bit larger, but as it stands, there was absolutely no reason to be using vitamin C prior to this publication, and this paper makes it pretty clear we shouldn’t be neutral on this issue, but instead suggest against it.
The metabolic cocktail saga is a fantastic cautionary tale that should remain at the front of our minds at all times in emergency medicine. It is all too easy for a little optimism and confirmation bias to cloud personal experience. There is a reason we rely on RCTs. There are thousands of exciting interventions – from REBOA to balanced transfusion – that we have more readily adopted, but which are similarly based on shaky evidence. (As much as I think replication is important in research, I find it sort of sad how many resources have been dumped into this topic, given the myriad of seemingly more important topics that require high quality research.)
Bottom line: Sepsis is not treated with vitamins. Be wary of low quality before and after research.
Is apixaban better than rivaroxaban?
Dawwas GK, Leonard CE, Lewis JD, Cuker A. Risk for Recurrent Venous Thromboembolism and Bleeding With Apixaban Compared With Rivaroxaban: An Analysis of Real-World Data. Ann Intern Med. 2022 Jan;175(1):20-28. doi: 10.7326/M21-0717. Epub 2021 Dec 7. PMID: 34871048
To date, there are no head to head RCTs comparing apixaban and rivaroxaban, and I think most physicians consider them to be interchangeable. This is a large database study that identified 50,000 patients diagnosed with VTE and treated with either apixaban or rivaroxaban. The headline news from this trial was that apixaban was associated with a lower rate of recurrent VTE (HR 0.77 95% CI 0.69-0.87) and also a lower rate of bleeding (HR 0.60 95% CI 0.53-0.69). When I first heard these results in abstract form, they sounded remarkable, because I couldn’t think of any reason there would be significant confounding between these two groups. (They also sounded rather unbelievable, as it would be pretty rare to get both less clotting and less bleeding from the same drug.) However, on closer reading, this result is only after propensity matching that eliminated 10,000 patients from the apixaban group and 3000 from the rivaroxaban group. In my mind, I would have expected the patients who received these two drugs to look very similar, but that isn’t the case. The apixaban group was older, and had a higher prevalence of renal failure, heart failure, hypertension, and diabetes. Aside from local formulary issues, I didn’t think there was any reason to choose one medication over the other. If anyone can explain to me why apixaban seems to be selectively used in older, sicker patients, I would really appreciate it. Ultimately, this apparent confounding makes me much less confident in these results, despite the authors’ valiant efforts to adjust for it. Database data is always messy. I thought this data might be closer to ‘random’, and therefore more reliable than some observational data, but I was clearly wrong. At best, this is hypothesis generating. There is an RCT underway looking at this exact issue, so until the COBRRA study is published, I will continue to practice the way I always have. That being said, seeing as there is no other reason to pick one agent over the other, it may be reasonable to choose apixaban preferentially until the RCT is published.
Bottom line: This observational data suggests apixaban may be both more effective and safer than rivaroxaban, but there are so many limitations when trying to adjust retrospective data from a large database, confidence in this finding should be very low.
I guess this shouldn’t be surprising: Just having an IV in place for an hour elevates your D-dimer
Heffner A, Kline J. Role of the peripheral intravenous catheter in false-positive D-dimer testing. Acad Emerg Med. 2001 Feb;8(2):103-6. doi: 10.1111/j.1553-2712.2001.tb01272.x. PMID: 11157283
Did everyone know that placing an IV can elevate D-dimer levels and increase false positives? Apparently we should, as this paper is more than 20 years old. This study included 20 healthy volunteers who had no risk factors for thromboembolism. They placed a 20 gauge IV in the antecubital fossa of all participants, took blood through that IV immediately to count as the baseline, and then 90 minutes later they performed a blood draw from the opposite arm. There was a statistical increase in D-dimer after the IV had been in place for 90 minutes, but the difference is miniscule (15 ng/mL vs 33ng/mL, p=0.04) and so I would probably have considered it clinically insignificant. However, the difference in false positives really stood out to me. There were 2 (10%) false positives on the immediate draw, but 5 (25%) on the blood draw at 90 minutes. Although this difference was not statistically significant, it certainly seems like it could be clinically significant. Of note, D-dimer values were probably artificially low in this study, because in addition to looking at young healthy volunteers, they handled the specimens with a perfect protocol (storing them on ice) that isn’t routinely followed in clinical practice. I don’t know if anyone out there applies this information clinically, or how you would. I haven’t seen any follow-up data. I was sort of hoping Jeff would respond on Twitter and just tell me how to interpret this study, but it hasn’t happened yet.
Bottom line: This is a very small study using only healthy volunteers, so take it with a grain of salt. However, it probably makes sense to try to ensure your DDimer is taken on the first blood draw, rather than after the patient has had an IV in place for hours in the ED.
One more PE study: Do subsegmental PEs matter?
Le Gal G, Kovacs MJ, Bertoletti L, et al. Risk for Recurrent Venous Thromboembolism in Patients With Subsegmental Pulmonary Embolism Managed Without Anticoagulation : A Multicenter Prospective Cohort Study. Ann Intern Med. 2022 Jan;175(1):29-35. doi: 10.7326/M21-2981. Epub 2021 Nov 23. PMID: 34807722
We know that overdiagnosis is a major problem in pulmonary embolism. We are finding too many, and hurting our patients in the process. However, it isn’t clear how to solve this problem. One hypothesis has been that subsegmental PEs are probably incidental, or overdiagnosis, and we might want to ignore them. This observational dataset of 292 patients in Canada and Europe looked at patients with subsegmental PEs referred to a thrombosis clinic and not treated with anticoagulation. The recurrent VTE rate at 3 months was higher than expected (3.1%), and may indicate that these patients would benefit from anticoagulation. (We can’t say that for sure. We know there are harms from treatment, and this is just this recurrence rate of PEs and DVTs found on imaging, rather than true clinical harms.) There are also reasons to think that this 3.1% number could be biased. These patients were all “carefully instructed on the signs and symptoms of venous thromboembolism and were instructed to contact study personnel if they occurred.” Furthermore, they were contacted at least 4 times during the 90 day follow-up period to be assessed for symptoms. Therefore, it is possible that many of the VTEs that occurred in follow-up were themselves overdiagnosed. There is irrefutable evidence that we are overdiagnosing PE, but I am not sure that the size of the PE is the factor that matters. My guess is that the underlying risk factors are far more important in deciding which patients need treatment, and which don’t. It is unclear based on these numbers whether patients with subsegmental PE require treatment, and an RCT is clearly warranted. You can read a full appraisal of this paper here.
Bottom line: The risk of recurrent VTE in subsegmental PEs may not be negligible. As we gather more data on the topic, the best approach for emergency physicians is to ensure that you are only testing for PE in high risk patients.
SciFi paper of the month: Bedside genomics
McDermott JH, Mahaveer A, James RA, Booth N, et al; PALOH Study Team. Rapid Point-of-Care Genotyping to Avoid Aminoglycoside-Induced Ototoxicity in Neonatal Intensive Care. JAMA Pediatr. 2022 May 1;176(5):486-492. doi: 10.1001/jamapediatrics.2022.0187. PMID: 35311942
Point of care genetic testing to guide therapy? It seems like science fiction, and is almost certainly not coming to your emergency department any time soon, but this study was too fascinating to pass on. As background, that I certainly did not know, there is a genetic variant (m.1555A>G) in the mitochondrial genome that substantially increases the risk of aminoglycoside-induced ototoxicity. This is a NICU based study, which makes some sense, as most neonatal antibiotic regimens include gentamicin (whereas I almost never prescribe gent in the ED). All neonates admitted to the NICU had a buccal swab done, which was run through a rapid-point of care gene drive system, and used to avoid gentamicin when appropriate in children with this mutation. The genotype only took an average of 26 minutes to do, so aside from cost, this is actually a practical test at this point. Of 751 neonates admitted, 424 (80%) were given an aminoglycoside. They identified 3 children with the genetic mutation, but there were another 5 false positives on the rapid test (based on confirmatory genetic testing), so the rapid genetic testing was not perfect. The study also wasn’t perfect, as 20% of children never had this genetic screening done, for a variety of reasons. There was no change in the time it took to administer antibiotics. Honestly, I have no idea how many genetic markers have been identified that would actually be useful to know in the ED. Given our history with testing, it is possible that this is far more hype than true benefit, with potential costs being astounding. On the other hand, there are so many potential uses, that this could represent a huge leap forward in medical care. We will want to see proper trials demonstrating true patient benefit, but if we can get genetic test results back faster than our basic metabolic panel, it is possible that I will see this technology in an emergency department during my career.
Bottom line: I include this for interest’s sake only, as it sounds more like science fiction than medicine at this point (at least based on the funding of the emergency departments I work in), but it will be an interesting field to watch over the coming years.
The title of this paper says it all: “Oral Is the New IV”
Wald-Dickler N, Holtom PD, Phillips MC, Centor RM, Lee RA, Baden R, Spellberg B. Oral Is the New IV. Challenging Decades of Blood and Bone Infection Dogma: A Systematic Review. Am J Med. 2022 Mar;135(3):369-379.e1. doi: 10.1016/j.amjmed.2021.10.007. Epub 2021 Oct 27. PMID: 34715060
If you have been following along for any length of time, you will know that oral antibiotics are clearly equivalent to (and possibly superior to) IV antibiotics in every single RCT ever done on the topic. I have spent a long time searching the literature, and I have not found a single study in which IV therapy beats oral (and yet I still see people returning to the ED almost every shift for their repeat IV antibiotics). This is a systematic review looking at the infections generally considered to need long courses of IV antibiotics: osteomyelitis, bacteremia, and infective endocarditis. They found a total of 21 RCTs, and “the 21 studies demonstrated either no difference in clinical efficacy, or superiority of oral versus IV-only antimicrobial therapy, including for mortality; in no study was IV-only treatment superior in efficacy. The frequency of catheter-related adverse events and duration of inpatient hospitalization were both greater in IV-only groups.” In other words, IV antibiotics were overall harmful, as compared to oral antibiotics. Oral treatment was actually statistically superior to IV treatment in endocarditis. Although there are problems with the underlying RCTs, the result should be obvious. As long as you have a functioning GI tract, the only possible physiologic difference between oral and IV antibiotics is the 30 minute absorption time for the very first dose. If the patient is going to the ICU, use an IV. Otherwise, and we please stop this harmful practice of putting patients on unnecessary IV antibiotics?
Bottom line: Even in the worst infections, there is no reason to choose IV antibiotics over oral.
Yes, I said “quiet”
Geller JE, Strickland PO, Bucher JT. The use of the word “quiet” in the emergency department is not associated with patient volume: A randomized controlled trial. Am J Emerg Med. 2022 Mar 16;56:10-12. doi: 10.1016/j.ajem.2022.03.020. Epub ahead of print. PMID: 35339973
If you publish a paper about saying “quiet”, you can be sure I am going to cover it. Usually, the topic alone would be enough to earn a stand alone post on the blog. However, despite covering a great topic, this paper sort of sucks, so I figured I would just cover it briefly here. I have seen a number of people report that this RCT demonstrates that saying quiet has no impact on the emergency department, but this paper doesn’t ask that question at all. They don’t present any information about how busy the department was; they only ask how busy people felt the department was. In the intervention grop, a researcher would approach as many staff members as possible and ask, “has it been quiet in here?” As the control, the same researcher would approach as many people as possible, but just greet them without saying the word quiet. Their outcomes were three survey questions about the staff member’s perception of the department 3 hours after the intervention. For some reason, they don’t tell us how many patients were seen on the quiet versus the non-quiet days (the most important piece of information), but they do say that “the use of the word ‘quiet’ was not associated with increased patient volumes”. There were not any statistically significant differences in their primary survey questions:
- “Where would you say is your feeling towards how crowded the emergency department is?”: 60/100 when quiet was said versus 54/100 when not said (p=0.27)
- “Where would you say is your feeling towards how many patients have entered the ER over the past few hours?”: 59/100 versus 52/100 (p=0.24)
- Where on this scale is your belief that the use of the word ‘quiet’ subsequently plays a role in how busy the rest of the day gets?”: 51/100 versus 56/100 (p=0.32)
Unsurprisingly, one’s answer to the first two questions was strongly associated with one’s answer to the third. (If you believe saying quiet plays a role in how busy the day gets, you assess the day as getting busier.) Looking at how your beliefs influence your perceptions is an interesting line of research, but it is somewhat pointless with providing us with (presumably easy to get) objective numbers about how busy it actually was in the department. This paper tells us absolutely nothing about the pervasive myth of magical words aside from the fact that a very large proportion of presumably smart and well trained people believe it.
If you are interested, there are numerous other RCTs that have already demonstrated what should be eminently obvious: saying quiet has absolutely no effect on how busy the emergency department is.
Bottom line: In the name of critical thinking education, I will continue to discuss how quiet my department is every single shift (and risk physical disability and lost friendships in the process)
DO NOT DELAY URGENT CONTRAST CTs WAITING FOR BLOOD WORK!!
Macdonald DB, Hurrell C, Costa AF, McInnes MDF, O’Malley ME, Barrett B, Brown PA, Clark EG, Hadjivassiliou A, Kirkpatrick IDC, Rempel JL, Jeon PM, Hiremath S. Canadian Association of Radiologists Guidance on Contrast Associated Acute Kidney Injury. Can Assoc Radiol J. 2022 May 24:8465371221083970. doi: 10.1177/08465371221083970. Epub ahead of print. PMID: 35608223 [free full text]
DO NOT DELAY URGENT CONTRAST CTs WAITING FOR BLOOD WORK!! That is probably the most important part of these guidelines. They do a good job discussing the evidence for this topic, and I think this short document is a worthwhile read for any emergency doctor. However, they do make one critical error. They fabricate the scenario we all warn our students against, in which we order tests that do not affect our management. If a contrast scan is required to make an important diagnosis, and there is no alternative diagnostic test, they suggest proceeding with the scan whatever the patient’s eGFR, but for some reason they still suggest checking the eGFR. That doesn’t make any sense. If you are going to proceed with the contrast CT whatever the blood work shows, waiting for the results of the blood work only delays the test and harms your patient.
Bottom line: These radiology guidelines tell us that we should not delay urgent contrast CT scans just to see renal blood work. Good luck getting that implemented.
Ordering tests just for reassurance doesn’t work
Rolfe A, Burton C. Reassurance after diagnostic testing with a low pretest probability of serious disease: systematic review and meta-analysis. JAMA Intern Med. 2013 Mar 25;173(6):407-16. doi: 10.1001/jamainternmed.2013.2762. PMID: 23440131
Does ordering tests help reassure patients with low risk presentations? I think this is a very important paper, but none of the studies involved emergency department patients, and there may be reasons to think that our population is different. This is a systematic review and meta-analysis that found 14 RCTs that compared testing to not testing in low risk presentations, and focused on the psychological impacts of those tests. More than half the studies looked at endoscopy in low risk dyspepsia, with other studies looked at imaging in back pain, MRI in chronic headache, event recorders in palpitations, and ECG and blood work in chest pain. The trials were small, with a total of 3828 patients across the 14 trials, and at significant risk of bias (none were blinded). However, the lack of blinding is sort of the point here. We assumed all these tests were going to be negative. We wanted to know if seeing the negative result helps alleviate patient anxiety about their symptoms. The quick answer seems to be no. There was no change in “illness concern” or subsequent symptoms in follow-up. Short term anxiety was unchanged. Long term anxiety actually comes very close to being worse with testing, but only 2 studies looked at this outcome, and the trend is driven entirely by one study looking at loop recorders in palpitations. (I feel like this meta-analysis must be missing studies, because I know I have seen studies demonstrating increased anxiety and decreased quality of life after MRI for low risk back pain. Of course, this is a paper from 2013, so I might be thinking of papers published since then.) They did find a statistical decrease in the total number of primary care visits after negative testing, which certainly could be a benefit, unless these patients have simply given up on that physician and sought care elsewhere, or actually have anxiety that the family doctor could be helping them manage. A systematic review can only be as good as the papers included, and there are some issues here, so we should be cautious in our interpretations. That being said, considering how frequently even very well meaning doctors order tests just for ‘patient reassurance’, I think it is important to know that the best available evidence seems to suggest that our tests don’t actually reassure our patients.
I think there is another important point to consider for emergency physicians: any potential benefit of reassurance will always be less when addressing acute conditions. The above studies focused primarily on patients with chronic abdominal pain. We expect those symptoms to continue for years, if not decades, and therefore I could imagine that reassurance could have a large benefit (though the studies didn’t show one). Instead, imagine you were ordering a CT head in a patient with a very low risk head injury for ‘reassurance’. The potential psychological benefit is so much smaller, because within 12 hours it will be very clear to the patient that their trauma caused no harm.
Bottom line: It is our job to reassure our patients with good education and time spent at the bedside. Tests can’t do that job for us.
Marketing-based versus evidence-based treatment in trigeminal neuralgia
Keppel Hesselink JM, Schatman ME. Phenytoin and carbamazepine in trigeminal neuralgia: marketing-based versus evidence-based treatment. J Pain Res. 2017 Jul 17;10:1663-1666. doi: 10.2147/JPR.S141896. PMID: 28761370
I was taught, and have always taught in turn, that carbamazepine is the first line treatment for trigeminal neuralgia. I also have a history of talking about the incredible harm done to our patients by the greed of drug companies. In that context, a paper entitled “Phenytoin and carbamazepine in trigeminal neuralgia: marketing-based versus evidence-based treatment” was bound to catch my eye. It is a great editorial covering the history of management of trigeminal neuralgia, and stands as a good reminder of the shaky scientific footings our profession came up on. The initial adoption of antiepileptics for trigeminal neuralgia was based on studies that look identical to the Marik paper on vitamin C in sepsis. A few very keen researchers published case report data on a handful of patients with claims that their novel therapy never failed. Unfortunately, in the 1950s, instead of following those case reports with RCTs, these therapies instead were adopted as ‘standard of care’ and retain that label to this day. However, that is not the main point of this editorial. Initially trigeminal neuralgia was treated with phenytoin, and by all reports, it was treated effectively. However, in the 1950s a new drug (carbamazepine) was invented, and it was pushed as the new standard without any science at all. Phenytoin was off patent already at that point, so carbamazepine was much more expensive. This is a pattern we see over and over again in medicine: new expensive drugs replace old effective drugs with absolutely no evidence. The authors summarize this wonderfully: “From an ethical perspective, we struggle with this. To substitute patent protection and aggressive marketing for scientific evidence of superiority or even non-inferiority potentially puts our patients at risk for oligoanalgesia in the name of corporate profits. Although we have highlighted the unfortunate histories of phenytoin and carbamazepine, these are certainly not the only pain medications with regard to which the wool has been pulled over the eyes of unsuspecting prescribers and their patients. Our hope is that clinicians will accept this editorial as a warning to familiarize themselves with the empirical data supporting the prescription of one drug over another, as our cautionary tale demonstrates that “the next great thing” is not necessarily all that great.”
Bottom line: The evidence for carbamazepine in trigeminal neuralgia is incredibly weak. You could probably treat it exactly like any other neuropathic pain syndrome. More importantly, physicians need to stop letting pharmaceutical companies take advantage of our patients by marketing more expensive treatments that have absolutely no evidence of benefit.
In fact, since writing the original summary, I found a meta-analysis (Yuan 2016 PMID 26891784) comparing gabapentin and carbamazepine in trigeminal neuralgia. It turns out that gabapentin is probably more effective, is preferred by patients, and definitely has fewer side effects. I might include a full write up in the future, but this sounds like a practice changer. (That being said, there has been evidence of significant publication bias in the evidence around gabapentin for other topics, so this review probably deserves a full critical read.)
COVID guidelines mostly suck
Zhao S, Lu S, Wu S, et al. Analysis of COVID-19 Guideline Quality and Change of Recommendations: A Systematic Review Health Data Science. 2021; 2021:1-22. https://doi.org/10.34133/2021/9806173
I hate guidelines. This COVID paper helps to explain why. They did a search for any COVID treatment guidelines published in 2020, and assessed their quality. (They excluded guidelines looking at disease prevention, which I expect would be even lower quality, given that treatment guidelines generally will be based on RCTs, whereas the IPAC and prevention guidelines are generally not.) Of the 92 guidelines that met their inclusion criteria, only 7 (8%) met the Institute of Medicine definition of a quality guideline, meaning that their recommendations were “informed by a systematic review of the evidence”. Think about that for a moment. That means that 92% of available COVID treatment guidelines were not informed by a systematic review of the evidence. And that is a pretty low bar, as that standard says nothing about whether the systematic review or guideline process was done well. Only 22% of these guidelines reported managing conflicts of interest. 73% of guidelines had no system to rate the quality of evidence or strength of recommendation, and 72% didn’t even say how consensus for recommendations were reached. Using a formal AGREE II tool for grading guidelines, the mean score was only 30%. In other words, these guidelines don’t even come close to getting a passing grade. (Horrid COVID guidelines were also discussed in this post.)
I don’t have space in the rapid round-up for the real rant that is required about COVID guidelines, but as they currently stand, COVID therapeutics guidelines will ruin emergency medicine. Guidelines that assume every patient with respiratory illness needs to be tested early, based on the assumption that we need to identify patients for outpatient treatments, none of which has high quality evidence, mean that a tsunami of patients will need testing in the coming flu season. In countries like Canada, with no plan for that testing, the obvious foreseeable disaster is that all of these patients will be sent to the emergency department.
Bottom line: This is about way more than COVID. Most guidelines are awful. You must read guidelines like any other published paper: with a very critical eye, focused primarily on methodology. Most guidelines lack any rigor, and should be ignored.
I don’t even know
Bush SP, Naftel J. Injection of a whole black widow spider. Ann Emerg Med. 1996 Apr;27(4):532-3. doi: 10.1016/s0196-0644(96)70256-1. PMID: 8604882
“A 37-year-old woman with a history of IV heroin use crushed a whole black widow spider, mixed it in 10 mL of distilled water, and injected the mixture intravenously.” Not actually a ton to learn from this 26 year old letter to the editor, but when I run across a letter like that, it is going to get included. In addition to severe muscle pain, the patient developed a severe asthma-like reaction, which could have been allergic in nature, or due to the direct smooth muscle constriction from the black widow venom. Either way, she did well after a day in the ICU.
Bottom line: There is no presentation too far fetched for emergency medicine.
Cheesy joke of the month
A police officer came to my door the other day.
He told me he was looking for a man with one eye.
I told him it would probably go faster if he used both.