This month’s articles are a little special. Usually, I pick out the articles that catch my eye, but it’s easy to imagine that my biases filter out a lot of interesting medical literature. As there were a number of experts descending on Berlin this month for the SMACC conference, Casey Parker and I invited a few people to suggest a paper and discuss it with us over a beer for our podcast. The conversation was great, and can be found here. These are the articles that were picked:
Sometimes EBM just makes me want to puke
Suggested by Dr. Pik Mukherji (@ercowboy) Egerton-Warburton D et al. Antiemetic use for nausea and vomiting in adult emergency department patients: randomized controlled trial comparing ondansetron, metoclopramide, and placebo. Annals of emergency medicine. 2014; 64(5):526-532.e1. PMID: 24818542
This is a prospective, randomized, double-blind, placebo controlled trial. In 270 patients in the emergency department with nausea and vomiting, they compared 4 mg of IV ondansetron, to 20 mg of IV metoclopramide, to placebo. (Neither of those are standard doses where I work.) The primary outcome was the sensation of nausea, as judged on a visual analogue scale. The simple answer is that there was no difference. I have some problems with the math they present, but the differences I calculated were a improvement of 33mm with ondansetron, 32mm with metoclopramide, and 25mm with placebo. (So placebo does look worse, but not statistically so, and probably not clinically so.) Unfortunately, there were so few episodes of vomiting that they can’t actually compare. There were a significant number of side effects in the metoclopramide dose, which isn’t uncommon at 20mg. These results shouldn’t be surprising. We have seen a few studies like this before. In fact, the only medication with good evidence for nausea and vomiting in adults in the emergency department is droperidol. However, these patients only started with nausea rated 5/10 and almost none of them vomited, so it might just be that they were treating the wrong patients. The patient population here will be very heterogenous, because there are so many causes of nausea and vomiting. Furthermore, there is excellent evidence that ondansetron limits vomiting in pediatric populations, and there is no reason to think that adults are substantially different. We probably overuse these drugs, but to address that question, we need to explore which subgroups are helped and which aren’t, rather than going after the class of medications as a whole. Also, if you aren’t going to use them, what are you going to use instead?
Bottom line: Antiemetics probably aren’t as useful as we assume or hope, but it isn’t clear what we should do instead, unless you have access to droperidol.
How to cause an arrest in pigs
Suggested by Andy Neill (@AndyNeill): Aagaard R et al. Detection of Pulmonary Embolism During Cardiac Arrest-Ultrasonographic Findings Should Be Interpreted With Caution. Critical care medicine. 2017; 45(7):e695-e702. PMID: 28403120
I generally try to stay away from animal studies here, but it is difficult to control cardiac arrest in humans. This is a study in pigs, examining the cardiac ultrasound in 3 different causes of arrest: PE, hypoxia, and primary arrhythmia. The most fascinating part of this study is how the simulated PE in pigs. They inserted an endotracheal tube into the pig’s left external jugular vein and then premade 5x1cm clots were sent down this tube. The reason that this paper was picked was that there was some degree of RV dilation in all the groups, raising the question of how accurate we can be diagnosing PE in this context. The PEs did have the most RV dilation. Clinicians were able to differentiate PE with a sensitivity of 79% and a specificity of 68%. Honestly, I am not sure what to do with this data. It is pigs after all. And we have seen previous studies where PEs were identified by ultrasound during cardiac arrest, and the patients had good outcomes after lytics. The real question is how do the benefits in the true positive group weigh against the harms in the false positive group. Given the dismal survival in PEA arrest, it is hard to image there is a significant harm of giving thrombolytics if you think there might be a PE.
Bottom line: Intra-arrest cardiac arrest ultrasound is not perfect for diagnosing PE (of course), but I think it should still be used to search for potential causes of arrest, and you should not hesitate to give thrombolytics if you suspect the arrest was caused by a PE.
Saturday night palsy? Don’t tell Saturn
This is a fun read. We have all heard of “Saturday night palsy”, the radial nerve palsy that is supposed to be named after the nerve compression that occurs when you pass out drunk in an awkward position on a hard public bench after a Saturday night spent out drinking. These authors explore the history a little bit, and suggest that the name may actually be the result of a historical game of broken telephone. “Saturnine palsy” refers the nerve palsy, primary affecting the radial nerve (usually bilaterally), that results from lead poisoning. Apparently lead is very sweet tasting, so people used to chew on their lead goblets. Anyway, they wonder whether, given the similarities in clinical presentation, “saturnine palsy” was bastardized into “Saturday night palsy”. After all, those of us working in emergency medicine know that people enjoy their alcohol every night of the week, not just Saturday.
Bottom line: There really isn’t one – just an interesting read.
Dead, or just mostly dead?
Suggested by Ben Smith (@UltrasoundJelly): Gaspari R, Weekes A, Adhikari S. Emergency department point-of-care ultrasound in out-of-hospital and in-ED cardiac arrest. Resuscitation. 2016; 109:33-39. PMID: 27693280
We heard above that intra-arrest cardiac ultrasound might not be perfect for assessing right ventricular dilation. This paper looks at whether cardiac activity on ultrasound is associated with improved survival. It is a prospective, multi-center, observational study that included a total of 793 non-traumatic cardiac arrest patients, in PEA or asystole, who had ultrasound performed during their arrest (so a select group of patients). 208 patients (26%) had ROSC, 114 (14%) survived to hospital admission, and 14 (1.6%) survived to hospital discharge. ⅓ of patients had cardiac activity on the initial ultrasound. Treating physicians were not blinded to ultrasound results, which is the biggest weakness of this data. (A lack of cardiac activity might have caused death, by limiting resuscitation, rather than predicting death). 3.8% of patients with cardiac activity versus only 0.6% of patients without cardiac activity survived to hospital discharge. (It was 29% vs 7% for survival to hospital admission). These are statistically and clinically different, but it isn’t clear that the difference was large enough to actually help you make a clinical decision, especially considering the biases in this unblinded data. Ultrasound did appear to help in some other ways. 34 patients had a pericardial effusion identified and pericardiocentesis was attempted in 13 of them. Survival to hospital discharge was much better than average: 15.4%. Similarly, there were 15 patients who were given thrombolytics for presumed PE, and 1 (6.7%) survived to hospital discharge. Another pearl: 10% of patients with asystole on the monitor had cardiac activity on ultrasound. My take away from all this? Intra-arrest cardiac ultrasound is not a perfect test. It isn’t clear exactly how it should be integrated into our algorithms. However, there is an association between cardiac activity and survival, and ultrasound will find some treatable causes of PEA arrest, so I think there is clearly a role for it. Just remember, it is not definitive. One other point: the survival rates are very low, even if there was ongoing cardiac activity on ultrasound. When running a code with the family in the room, it can be very difficult to explain the decision to stop a resuscitation when the fancy technology that everyone can see still shows a beating heart. (This can also sometimes be hard to explain to team members.) However, this data clearly indicates that survival from PEA and asystole is still very low, even in the presence of cardiac activity, and we still need to make a clinical decision.
Bottom line: Like all ultrasound use, intra-arrest cardiac ultrasound is an important tool, but is not definitive in isolation, and must be interpreted within the clinical context.
Memories of van Gogh
Suggested by Michelle Johnston (@Eleytherius): Chan BS et al. Efficacy and effectiveness of anti-digoxin antibodies in chronic digoxin poisonings from the DORA study (ATOM-1). Clinical toxicology (Philadelphia, Pa.). 2016; 54(6):488-94. PMID: 27118413
Acute and chronic digoxin poisonings are very different clinical entities. Digoxin has a very large volume of distribution, meaning that very little is in the circulation at one time. Because of that, these authors hypothesized that, in chronic toxicity, the use of anti-digoxin antibodies might cause lower free digoxin levels without significantly impacting clinical outcomes. (I wonder, instead, if this tells us that Digibind should be given as a slow infusion rather than as a bolus?) This is a prospective observational study of 36 patients with chronic digoxin toxicity (a level >2.6nmol/L plus symptoms). Looking at the baseline characteristics, I wonder if all these patients really needed to be treated in the first place. (Median heart rate was 49. A lot of these patients might have been treated conservatively where I work.) Serum levels definitely went down after the Fab was given. Heart rate response was not as vigorous, although there was a dose-response relationship, with an increase by 4.5 beats/min in patients given 1 vial, 10 bpm with 2 vials, and 17.3 bpm with 3 vials. The authors say this is only a moderate response, but a change of 17 in heart seems clinically relevant to me. Less important, but I will also note that all 22 patients with GI symptoms saw those symptoms relieved. Another pearl: levels did fall to almost 0 during administration, but rebounded above 2nmol/L in 25 patients.
Bottom line: Digoxin antibodies may not be a panacea, but contrary to these authors, I think this data does show some clinically important results. My takeaway is more that we don’t know how to dose Digibind, nor when to give it.
You say pink eye, I say red eye
Suggest by Dr. Connie Smith (@halfbakeddoc) Rose PW, Harnden A, Brueggemann AB. Chloramphenicol treatment for acute infective conjunctivitis in children in primary care: a randomised double-blind placebo-controlled trial. Lancet (London, England). 2005; 366(9479):37-43. PMID: 15993231
This is a randomized, double blind, placebo-controlled trial that included 326 children with a clinical diagnosis of conjunctivitis, comparing chloramphenicol drops to placebo. First, I should note that the placebo drops were not saline, but included both boric acid and borax, both of which are used as antiseptics, which doesn’t sound like the ideal placebo to me. Chloramphenicol was chosen because it is the most commonly chosen drop locally and has a low rate of resistance. Only 29% of eligible children were actually recruited to the study (possible selection bias). They did swabs, and in this study 80% of cases were bacterial, which sounds high to me. Using clinical cure as the primary outcome, at 7 days 86% of the chloramphenicol group was cured as compared to 79% of placebo. The difference was exactly the same if you just looked at the patients with a bacterial cause. The main take home is that most children will get better no matter what you do. 20% might still have symptoms at a week, and with treatment you might be able to drop that as low as 15%. Small trials like this don’t give us a great sense of harms, so knowing the harm benefit balance of this modest clinical impact in a mostly self resolving condition is difficult.
Bottom line: Most kids with conjunctivitis don’t need treatment, but there is a modest benefit from topical antibiotics.
Apneic oxygen, but now for kids
Suggested by Andy Tagg (@andrewjtagg): Humphreys S et al. Transnasal humidified rapid-insufflation ventilatory exchange (THRIVE) in children: a randomized controlled trial. British journal of anaesthesia. 2017; 118(2):232-238. PMID: 28100527
Andy suggested a paper on apneic oxygenation in pediatrics. They looked at 48 healthy children with normal airways and normal cardiorespiratory function (in other words, unlike emergency patients in every way) who were presenting for elective surgery. After inducing and paralyzing the children, they compared a basic jaw thrust with no oxygen or ventilatory support to a jaw thrust combined with high flow humidified oxygen. They report their outcomes by age range, but the average safe apnea time was almost doubled in the children receiving oxygen (up to 430 seconds with no ventilations and no desaturation). The biggest problem with this data is that they automatically stopped when they reached twice the expected safe apnea time, so no child in the apneic oxygenation group desaturated. (The average oxygen saturation was 99.6% at the end of the apnea period). There was a linear rise in carbon dioxide, and they point out that, because of pediatric metabolic rates, the rise in CO2 is about double that in adults. This might limit the safe apnea time, although no clinical consequences were seen here. It isn’t clear that the special humidified devices are necessary here, even if using relatively high flow rates, as we are doing so for a very short period of time. Again, these were not emergency patients, and the outcomes measured were not patient oriented. We already had this kind of data: you can keep people oxygenated for very long times despite apnea. What we are lacking is evidence that this improves any patient oriented outcomes in an emergency setting. That being said, it just makes sense, and while waiting for the evidence, I would rather give myself the extra time than not.
Bottom line: Nasal oxygenation clearly prolongs safe apnea time in healthy patients.
A bunch of papers on digital rectal massage – yes, really
Roberge R et al. Termination of paroxysmal supraventricular tachycardia by digital rectal massage. Annals of emergency medicine. 1987; 16(11):1291-3. PMID: 3662193
I am working backwards a little bit here. I was originally reading about digital rectal massage for the treatment of intractable hiccups, but one of those papers was a letter in response to this paper. This was a patient with SVT. They tried carotid massage twice, but that didn’t work, so they moved on to digital rectal massage “using a continuous circumferential motion.” This apparently resulted in an abrupt conversion to sinus rhythm – presumably because the patient was so disturbed at what was going on. I should also mention, the patient’s blood pressure was 70/50 while they were doing all this. If I saw this happening in my department, I am not sure it is the Annals of Emergency Medicine that I would be writing to. The authors suggest, however, that “this is a simple, rapid maneuver that may be preferable to other such techniques because of a decreased likelihood of complications noted with other vagotonic maneuvers.” It is unclear to me if any of these authors has ever actually spoken to a patient.
Lieberman ME. Ventricular tachycardia as a complication of digital rectal massage. Annals of emergency medicine. 1988; 17(8):872. PMID: 3395001
This is a letter to the editor in response to paper #1. This is another SVT patient, this time with a blood pressure of 108/70, where they tried Valsalva and carotid massage, and then moved on to digital rectal massage, again emphasizing the circumferential movement. This time, however, on the second circumferential movement, the patient converted into ventricular tachycardia. Woops. Luckily it was short lived, and the patient did subsequently convert to sinus rhythm.
Fesmire FM. Termination of intractable hiccups with digital rectal massage. Annals of emergency medicine. 1988; 17(8):872. PMID: 3395000
This is also a letter to the editor in response to paper #1. This is about a 27 year old man with 3 days of hiccups. Apparently a number of things had been tried, including other techniques I was not aware of like “tongue pulling” and gagging, but he wasn’t getting better. So, someone for the brilliant idea to try digital rectal massage, again in “a slow circumferential movement”. It was successful within 30 seconds, so who am I to judge?
Odeh M, Bassan H, Oliven A. Termination of intractable hiccups with digital rectal massage. Journal of internal medicine. 1990; 227(2):145-6. PMID: 2299306
We’ll finish up with only last case report on this technique: a 60 year old man who had hiccups caused by an NG tube. (Lesson number 1: NG tubes are rarely needed.) These guys also gagged the patient, in addition to trying “a teaspoon of granulated sugar”, carotid massage, and digital eyeball pressure. For the first time in the reports, the did try medications: metoclopramide, diazepam, and haloperidol. Ultimately, they lucked into the rectal technique. Someone performed a rectal exam, and the hiccups stopped. Unfortunately, they started again a few hours later, so this time the did the proper massage, with “a slow continuous circumferential motion” and the hiccups were cured forever. Please feel free to quote me by name when raising this therapeutic option with your patients.
Cheesy Joke of the Month
If you’re American in the living room what are you in the bathroom?
That’s it for this time. I can’t wait to do it again at the next SMACC conference in Sydney!
For more critical appraisal and “Articles of the month” summaries, click here.