Here is another collection of the articles I have found interesting in my reading from the last month or so. For this edition of the Research Roundup we have status epilepticus, nocebos, the pink lady, McGyver bias, and so much more. Podcast version over on BroomeDocs.
Our ongoing obsession with fever
Young PJ, Bellomo R, Bernard GR, et al. Fever control in critically ill adults. An individual patient data meta-analysis of randomised controlled trials. Intensive care medicine. 2019; 45(4):468-476. PMID: 30741326
This is a meta-analysis looking at individual patient data in RCTs looking at the question of whether fever management improves survival in critically ill patients. They identified 13 RCTs, and ultimately included 1413 patients in this analysis. One issue is that there were a variety of different techniques used to lower temperature, and there were also multiple different thresholds for treatment. This analysis looked at the “more active fever management” as compared to “less active”. The more active group did achieve a lower temperature, although the difference wasn’t huge (0.7 degrees – that’s Celsius of course – at 12 hours, and less than 0.5 degrees at all other time points). There was no difference in survival, ICU mortality, ICU length of stay, or hospital length of stay. I think this is consistent with everything we know about fever. Even though there is an association between fever and bad outcomes, there is no evidence that treating fever improves outcomes. (We should keep that in mind when discussing temperature management after cardiac arrest. It is frequently said that we “must avoid fever”, but we don’t have any trials demonstrating that treatment improves outcomes. There is only an association between fever and bad outcomes, which is not the same thing.)
Bottom line: Treating fever in critically ill patients does not improve survival.
The “it’s newer and more expensive so it must be better” bias
Dalziel SR, Borland ML, Furyk J, et al. Levetiracetam versus phenytoin for second-line treatment of convulsive status epilepticus in children (ConSEPT): an open-label, multicentre, randomised controlled trial. The Lancet. 2019; http://dx.doi.org/10.1016/S0140-6736(19)30722-6
Lyttle MD, Rainford NEA, Gamble C, et al. Levetiracetam versus phenytoin for second-line treatment of paediatric convulsive status epilepticus (EcLiPSE): a multicentre, open-label, randomised trial. The Lancet. 2019; http://dx.doi.org/10.1016/S0140-6736(19)30724-X
These are two similar RCTs looking at levetiracetam as the second line agent in pediatric status epilepticus (after benzodiazepines). Both trials were designed assuming that levetiracetam would be superior to phenytoin. ConSEPT included 234 patients at 13 hospitals in New Zealand and Australia. EcLiPSE included 286 patients from 30 hospital in the UK. The biggest difference was that ConSEPT excluded patients already on levetiracetam or phenytoin, whereas EcLiPSE didn’t. The results of both trials were the same: no difference. Levetiracetam was not better than phenytoin. Although some people have interpreted the “no difference” to mean that the drugs are equivalent, that is not what these trials showed. Superiority trials are no designed to show equivalence. In ConSEPT, levetiracetam resulted in 10% less (absolute) resolution of seizures, with a 95% confidence interval that extended down to 22% worse. These trials do not tell us it is equivalent, and in fact, it might be a lot worse. They do say that levetiracetam is not superior to phenytoin.
Bottom line: There trials do not provide any reason to replace phenytoin with levetiracetam in our status epilepticus algorithms.
The full write up can be found here.
Phenobarb as second line in status epilepticus
Burman RJ, Ackermann S, Shapson-Coe A, Ndondo A, Buys H, Wilmshurst JM. A Comparison of Parenteral Phenobarbital vs. Parenteral Phenytoin as Second-Line Management for Pediatric Convulsive Status Epilepticus in a Resource-Limited Setting. Front Neurol.. 2019; 10. [article]
I have long argued that our current algorithms for status epilepticus fail far too often, and that a more aggressive approach is required. I have suggested propofol, but the same logic would also suggest phenobarbital is a reasonable option. This is a single center RCT out of South Africa comparing phenobarb to phenytoin as the second line agent in status, and the results were pretty clear: in patients whose seizure didn’t respond to benzodiazepines, a single dose of phenobarbital worked 86% of the time, as compared to only 46% seizure cessation with phenytoin. People are always worried about respiratory depression with this aggressive approach, and it did occur in 56% of patients given phenobarbital. However, it occured in 70% of patients given phenytoin. The best treatment for respiratory depression is getting the seizures stopped as soon as possible. This is single center open label trial, and the etiologies of seizures in South Africa is somewhat different than I see, so there are significant limitations. However, I think it clearly adds to the evidence that our current approach is not good enough. I basically use the algorithm described here, except that I substitute propofol for phenobarbital because it is more widely available and I have more experience using it.
Bottom line: Phenobarbital might be a much better second line option for seizures than phenytoin (and therefore, based on there two trials above, also better than levetiracetam).
The full write up can be found here.
Can placebos kill? (If so, Dr Oz should be worried)
Reeves RR, Ladner ME, Hart RH, Burke RS. Nocebo effects with antidepressant clinical drug trial placebos. General hospital psychiatry. 2007; 29(3):275-7. PMID: 17484949
This is an interesting paper, suggested by Casey Parker, that might actually stir up some controversy when we discuss it for the podcast. It is a case report in which the authors describe a 26 year old male who was in a placebo controlled trial for an antidepressant and took an overdose. He presented pale and diaphoretic, with a heart rate of 110 and a blood pressure of 80/40. They describe caring for him, including giving 6 litres of IV fluid (if it was saline, some people are going to be really upset), but he was still lethargic with a blood pressure of 100/62 and a heart rate of 106. When they were finally able to contact a representative from the trial, and they found out he was on placebo, and he rapidly improved. Although the authors paint this as an amazing illustration of the nocebo effect (negative outcomes from placebos), I think we should be careful. This is not the most convincing case report in the world. It’s been a while since I did a deep dive on placebo, but my understanding of the science is that there is good evidence that placebo affects subjective experiences, but that is probably doesn’t have any real physiologic effects. Obviously thoughts can change physiology (think panic attack) and vice versa, but in general I think we overestimate the physiologic effects placebos have (and also probably underestimate their psychological effects). This this case in particular, the patient got better after learning he had taken placebo, but also after significant treatment. Was this really the nocebo effect, or did he improve because of the 6 litres of fluid and time? Did he only take placebo, or were there also co-ingestants? They describe respiration as being rapid, but don’t give any more details. Hyperventilation in itself can cause hypotension. Perhaps they just describing the physiologic effects of his anxiety (probably coupled with mild dehydration and malnutrition based on his baseline depressed state).
Bottom line: Placebo is a fascinating topic that I should probably review in more depth at some point, but I don’t think this case report is nearly as convincing for the “nocebo” effect as the authors make it out to be.
The “Pink Lady”: a bad placebo?
Berman DA, Porter RS, Graber M. The GI Cocktail is no more effective than plain liquid antacid: a randomized, double blind clinical trial. The Journal of emergency medicine. 2003; 25(3):239-44. PMID: 14585449
This is a classic RCT comparing an antacid (Mylanta) alone, to two “GI cocktails” which added an anticholinergic agent and viscous lidocaine to the antacid. The pain relief was the same in all groups. Thus, there is no benefit from the “pink lady” as compared to just using an antacid alone. Considering the extra harm involved in combining multiple drugs, it seems hard to justify that practice.
Bottom line: It’s time to retire the “pink lady”.
The full write up can be found here.
The MacGyver Bias
Duggan LV, Marshall SD, Scott J, Brindley PG, Grocott HP. The MacGyver bias and attraction of homemade devices in healthcare. Canadian journal of anaesthesia. 2019; PMID: 30980239
This is a great commentary which coins the term “MacGyver bias” to describe the inherent attraction to improvised (“MacGyvered”) devices in medicine. We all know many tricks for using standard medical equipment for unintended purposes. We know how to attach an angiocath to a bag valve mask through a 3 mL syringe. We can create a dental splint out of pieces of a facemask and some skin glue. We can stick a foley catheter into almost any hole, for a large variety of purposes. We use these solutions for a variety of reasons. Sometimes there no device that fills the role we need. More often, our hospital simply doesn’t stock what we need, or the equipment is missing, or stored in an inconvenient location. When we hear about these solutions, our first instinct is to praise the ingenuity of the inventor. But these authors want us to pause and think about the potential downsides. How often do these contraptions fail? How often do they cause harm? How often is a critical procedure delayed because the doctor was taught 3 different MacGyvered approaches to the problem and just can’t decide which to use, or confuses the necessary parts from 2 options so that the improvisation doesn’t actually work? We are very unlikely to know, because doctors are not rushing to publish that one time they tried to use a ballpoint pen for a cricothyroidotomy, but the pen broke and the patient died. Furthermore, medicine for the most part is predictable. What made MacGyver great is that he found himself in unpredictable situations and managed to improvise. However, if he knew in advance the problem he was going to face, and simply didn’t bring the appropriate equipment because he thought his paperclip was enough, we would think he was an idiot. The later scenario is probably a better description of most improvisation in medicine. MacGyvered solutions are not going to disappear in medicine, but rather than teach those solutions as a valid technique, we should probably use them to help us identify scenarios where our available equipment is suboptimal and work to remedy the underlying problem. (Although realistically this might require some changes to patent law when it comes to life saving equipment, because price is a large part of this problem).
Bottom line: When you see a MacGyvered solution in medicine, remember to consider the underlying system failures that made that improvisation necessary.
I really wish skin glued existed when I was a kid
Ste-Marie-Lestage C, Adler S, St-Jean G, et al. Complications following chin laceration reparation using tissue adhesive compared to suture in children. Injury. 2019; PMID: 30961924
This is a chart review looking at 2000 children with facial lacerations aiming to determine the difference in outcomes between sutures and skin adhesives. Skin adhesive was used in 89% of these patients. (As a side note, topical anesthetic was only used in 11% of patients, which doesn’t seem high enough to me). Being observational, the conclusions we draw will be limited. Overall, dehiscence (0.6%) and infection (0.1%) were rare, and there was no difference between sutures and glue. Focusing specifically on chin lacerations, which have a higher rate of dehiscence, there also wasn’t a statistical difference (2.2% vs 0%, absolute difference 2.2%, 95% CI -7.5% to 4.4%). This is a wide confidence interval, because of the small numbers, but this certainly could be consistent with glue being worse. On the other hand, every study I have reviewed to date has demonstrated no difference between sutures and glue, and this is a negative study.
Bottom line: Skin glue will still be my go to for most lacerations, even of the chin. However, in this trial physicians chose to use sutures in patients they thought were higher risk (for whatever reason), and I will also do the same.
The continuing tales of TXA as a wonder drug
Wand O, Guber E, Guber A, Epstein Shochet G, Israeli-Shani L, Shitrit D. Inhaled Tranexamic Acid for Hemoptysis Treatment: A Randomized Controlled Trial. Chest. 2018; 154(6):1379-1384. PMID: 30321510
This is a small, single center RCT looking at the use of nebulized TXA for patients with small volume hemoptysis who are admitted to the hospital. There are a large number of methodologic issues with this trial, which I explore more in the main post, which mean that I don’t have a lot of faith in the results of this trial. For what it’s worth, the TXA group seemed to do better across the board, with earlier resolution of bleeding, less total blood, and fewer procedures. Be careful applying these results in the emergency department, because massive hemoptysis was specifically excluded.
Bottom line: There may be some benefit from inhaled TXA in small volume hemoptysis, but this trial was too flawed to convince me that it should be standard care yet.
The full write up can be found here.
Don’t prescribe tramadol
Thiels CA, Habermann EB, Hooten WM, Jeffery MM. Chronic use of tramadol after acute pain episode: cohort study. BMJ (Clinical research ed.). 2019; 365:l1849. PMID: 31088782 [free full text]
The day after I published my anti-tramadol rant, this paper was published in the BMJ. It is a large (357,000 patients) retrospective database study looking at opioid use after surgery in Medicare patients in the United states. They looked at patients who filled an opioid prescription after one of 20 common operations, and divide them into 5 groups: no opioid fill, any long acting opioid (with or without any short acting opioid, including tramadol), tramadol only, a short acting opioid other than tramadol alone (reference group), or tramadol plus another short acting opioid. Patients had to be insured at least 90 days before surgery and 180 days after to be included in the sample. They excluded patients who were on opioids pre-operatively. The primary outcome they were looking at was prolonged opioid use, defined as use of opioids for 90 or more days starting in the 180 days after surgery. When compared to standard short acting opioids, tramadol use was associated with an increase in additional opioid prescriptions and persistent opioid use (47% increase in the adjusted risk; 95% CI 1.25 to 1.69; absolute increase 0.5%; P<0.001). Prescriptions for more pills initially and the use of long acting opioids were also associated with increased persistent opioid use. There are a number of limitations. Persistent opioid use doesn’t mean harm. The absolute numbers are small. And there may be hidden confounders (maybe providers are choosing tramadol in patients with a higher risk of opioid abuse). However, this data dispels the myth that tramadol is somehow safer (and fits with everything else we know about this drug).
A similar association was seen in another database study, in which tramadol was the highest risk short acting opioid, with increased persistent opioid use at both 1 and 3 years. At 1 year, 14% of the patients prescribed tramadol were still taking an opioid, as compared to 5-9% if the initial prescription was for another short acting opioid. (Shah 2017)
Bottom line: Once again, there is no reason to be prescribing tramadol.
Grape or grain but never the twain?
Köchling J, Geis B, Wirth S, Hensel KO. Grape or grain but never the twain? A randomized controlled multiarm matched-triplet crossover trial of beer and wine. Am J Clin Nutr 2019;109(2):345-352. PMID: 30753321
There are a variety of rhymes that tell you what order you should consume your alcohol in. I love that these authors felt compelled to study their veracity. In this RCT, one group drank beer first then wine, another group did the opposite, and a third group drank only one product. All participants drank to a blood alcohol level of 0.11% or higher. (If you were wondering, they served Carlsberg and Edelgräfler.) The next week, the groups swapped over and drank in the opposite order. There were no differences between the groups on a 56 point acute hangover scale. There were also no differences in the hangover scores between male and female participants. (They were drinking to a target blood alcohol level, not to a set number of drinks). This study was industry supported, with beer donated by Carlsberg. These were university volunteers, so the results might no extrapolate to more refined drinkers. Also, a blood alcohol of only 0.11% might not be enough to make a lot of people feel hungover. (The overall hangover scores were low – only about 15 out of 56). Furthermore, almost half the population was lost to follow-up over just 1 week, which doesn’t make a lot of sense to me in a study offering free alcohol to university students.
Maybe my biggest take away from this paper is that the medical term for hangover is “veisalgia”. How did I not know that?
Bottom line: Drink whatever you want, in whatever order you want, I guess.
If you come across articles that you think should be included in this roundup – especially the weirder “Ignoble prize” type – please send them my way. Thoughts or comments on these articles are welcome below.
Cheesy Joke of the Month
Why shouldn’t you use a broken pencil?
Because it’s pointless
Morgenstern, J. Research Roundup (May 2019), First10EM, June 10, 2019. Available at:
Great opportunity for residents
You have the chance to win a scholarship to the ResusX (www.ResusX.com) course September 30 to October 2 this year. This is a resuscitation and critical care course run by Haney Mallemat, with some of the best teachers in the world, including Scott Weingart, Evie Marcolini, Rory Spiegel, Anand Swaminathan, and Salim Rezaie. To win a scholarship (free registration and some additional perks):
- Residents should submit a 3-minute video on how to do a procedure, explaining a complex topic, etc.
- Videos will be reviewed and finalists will be selected.
- The finalists videos will be posted to social media for voting and the winner will be announced.
Submissions should be sent to firstname.lastname@example.org. Sounds like a great opportunity. Good luck!
For more nerdy medical education, be sure to subscribe:
2 thoughts on “Research Roundup (May 2019)”
The article, “Our Obsession with Fever” was interesting. I am a paramedic working toward getting into medical school so there is my current level of education. It is my understanding that when it comes to fighting off an infection, fever is a natural response. I can see where continuous high fever would be detrimental if the patient is nutrient deficient due to the increased metabolism. Could the association between fever and poor outcomes just be the severity of the disease process? The higher the fever, the more severe the infection? I plan to read more on this. Great site!
Yeah – there are lots of reasons that fever might be associated with bad outcomes. If it was as easy as increased metabolic demand, you might assume that lowering the temperature (and therefore lowering the metabolic demand) wold lead to improved outcomes. Seeing as that hasn’t been the case, I think you have to look elsewhere (and these hidden confounders are not always easy to identify). For example, patients with bad neurologic damage might have impaired temperature control, leading to higher temperatures, but it is actually the neurologic disease rather than the temperature that leads to ultimate bad outcome.