Welcome to the May 2016 edition of my favourite reads from the medical literature. This will probably be the last post on First10EM for a little while, as I plan to take a summer vacation as well as a prolonged Ireland stay for SMACC.
Gastro game changer
Freedman SB, Willan AR, Boutis K, Schuh S. Effect of Dilute Apple Juice and Preferred Fluids vs Electrolyte Maintenance Solution on Treatment Failure Among Children With Mild Gastroenteritis: A Randomized Clinical Trial. JAMA. 315(18):1966-74. 2016. PMID: 27131100
Kids just want to drink juice, but I’ve been told that if I let kids with gastro drink juice they will die (or something like that). This is a randomized, controlled non-inferiority trial out of the Hospital for Sick Children that compared an electrolyte solution to a combination of half strength apple juice in the ED and the child’s preferred fluid (juice or milk) at home. 647 children aged 6 to 60 months with acute (less than 96 hours) diarrhea or vomiting with mild dehydration were included. For the primary outcome, which unfortunately was a composite of a number of things including IV use, hospitalization, health care contact, and prolonged symptoms, the juice group had a ‘treatment failure’ rate of 16.7% as compared to 25.0% with the electrolyte solution (difference 8.3%; 97.5% CI 2% – infinity). Converting from a non-inferiority analysis to a superiority analysis resulted in a p value of 0.006. In other words, the juice group was statistically better than the electrolyte group. The biggest caveat is that these kids were not sick, so the results could be different in kids with even moderate dehydration.
Bottom line: I am no longer forcing kids to drink something they hate. Whatever their preferred liquid is, it will keep them hydrated.
This is another paper that will be featured on EMCases Journal Jam. If you have questions you want the author to answer, let me know.
I’m not so ENCHANTED
I am honestly surprised that I managed to read an entire article that started with the statement: “Thrombolytic therapy with intravenous alteplase (recombinant tissue-type plasminogen activator) at a dose of 0.9 mg per kilogram of body weight is an effective treatment for acute ischemic stroke, despite increasing the risk of intracerebral hemorrhage.” (This is a good reminder that when reading articles, it is often best to just skip the introduction. This section is just a non-systematic review of the topic, aka a statement of the author’s biases and opinions.) Moving beyond that, this was a multi-center, prospective, randomized, open-label, non-inferiority trial comparing usual dose tPA (0.9mg/kg) to low dose tPA (0.6mg/kg) in 3310 patient with acute ischemic stroke within 4.5 hours of onset. (As a reminder of how rarely we use this intervention, they screened 69305 patients to enroll those 3310.) The primary outcome was a composite of disability and death, defined as a modified rankin score of 2 or more. There was no statistically significant difference between the two groups (53.2% low dose and 51.1% usual dose, p=0.51). However, the 95% confidence intervals around this result go beyond a pre-specified definition, therefore they were unable to demonstrate non-inferiority. For the primary harm outcome, there was less intracranial bleeding in the low dose group (1.0% vs 2.1% (p=0.01) by SITS-MOST criteria or 5.9% vs 8.0% (p=0.02) by NINDS criteria). Death at 7 days was lower in the low dose group, but death at 90 days was unchanged. There are a number of problems with this study. I am not going to delve too deeply into the issues of comparing different doses of placebo. (A dose response relationship is generally something we look for in efficacious therapies.) There is really no reason to make a study like this open-label and unfortunately that introduces a number of potential biases. Almost all the follow up was done by phone and the modified Rankin score is notoriously unreliable. Finally, like almost all of this research, the authors have significant conflicts.
Bottom line: High dose, low dose, no dose? I like this line of research. Maybe we can just keep lowering the tPA dose until is diluted by a factor of 10300 and hand care of acute stroke patients over to the homeopaths.
Therapeutics Initiative. Benefits and harms of drugs for “neuropathic” pain. Therapeutics Letter. 2016; 96:1-2. [free full text]
We see a lot of chronic pain. More than a lot. Since I started practicing, the number of people on gabapentin or lyrica for their neuropathic pain has skyrocketed. But just how good are these medications? This therapeutics letter looks at the evidence summarized in 11 different Cochrane reviews, and the best evidence on the topic is:
- The evidence is weak (surprise anyone?) and the available RCTs have a high risk of bias
- At best, about 1/10 patients will achieve any meaningful reduction in pain
- Almost everyone has some side effects from these drugs
- If there is going to be a benefit, you will see in within about 1 week
- There does not seem to be any benefit in higher doses. (I think this is the most important takeaway, as I often see people on crazy escalating doses)
Bottom line: It might be reasonable to try these medicines, but start at a low dose, and recheck at 1 week if benefits outweigh side effects. If they don’t, stop the drug.
Black box on fluoroquinolones
FDA Drug Safety Communication: FDA advises restricting fluoroquinolone antibiotic use for certain uncomplicated infections; warns about disabling side effects that can occur together [available here]
So this isn’t research – it’s a notice of a new black box warning from the FDA. I am often skeptical of these warnings, as in some cases I think they have clearly done more harm than good (droperidol), but I think this one is worth knowing about. They say that “the serious side effects associated with fluoroquinolone antibacterial drugs generally outweigh the benefits for patients with sinusitis, bronchitis, and uncomplicated urinary tract infections who have other treatment options. For patients with these conditions, fluoroquinolone should be reserved for those who do not have alternative treatment options.” I would say this is pretty obvious, as there is no benefit of antibiotics in sinusitis and bronchitis. If a patient has a bad outcome and you are giving them a drug that has no chance of helping them, I would find that hard to defend.
Bottom line: Don’t use fluoroquinolones first in uncomplicated UTIs. Don’t use antibiotics at all in sinusitis or bronchitis.
But can we change our bad antibiotics habits?
Meeker D, Linder JA, Fox CR. Effect of Behavioral Interventions on Inappropriate Antibiotic Prescribing Among Primary Care Practices: A Randomized Clinical Trial. JAMA. 315(6):562-70. 2016. PMID: 26864410
We know that antibiotics don’t help for the vast majority of acute respiratory tract infections, but for some reason people just can’t help themselves. Every patient I see in the ED with a cough for 3 days is already on azithromycin or amoxicillin. They come to the ED because they can’t understand why they aren’t getting better on antibiotics. They think they need something stronger. This is a randomized controlled trial trying to get doctors to smarten up. In a total of 243 different clinicians, they tried three different interventions: 1) automated reminders that antibiotics are inappropriate and alternate treatment suggestions 2) the requirement of an ‘antibiotic justification note’ as part of the permanent record, and 3) intermittent e-mails comparing the performance of various doctors. Each clinician was exposed to anywhere from 0 to 3 of the interventions. The control group (no intervention) decreased their prescribing rate by 11% – a good demonstration of the Hawthorne effect. The reminders did nothing. Requiring a justification and being compared to peers decreased inappropriate antibiotic use.
Bottom line: Sadly, simply giving physicians information is not enough to change their practice. We need to be shamed into change. Maybe I should stop writing about the evidence and instead walk around personally shaming people?
Again – we don’t listen to good advice
Rosenberg A, Agiro A, Gottlieb M. Early Trends Among Seven Recommendations From the Choosing Wisely Campaign. JAMA internal medicine. 175(12):1913-20. 2015. PMID: 26457643
I love the choosing wisely campaign – except that we know doctors love to ignore good advice. This is a retrospective look at a billing database (so not necessarily the most reliable data, although the conclusions are believable.) They looked at 7 items that were listed by choosing wisely as being of minimal or no benefit (such as pre-op chest x-ray in the absence of concerning history, or imaging of low back pain without red flags) and looked to see if the number billed for changed over a 3 year period after the recommendations. They didn’t. (OK, imaging for headache went down from 14.9% to 13.4% – not exactly a clinically important change). Horrendously, the use of antibiotics for sinusitis remained at 84%!
Bottom line: Physicians just don’t change their practice when presented with good evidence or advice. It does makes me wonder if I should stop sending these e-mails – as they are probably not accomplishing anything.
Police officer: “Sir, How high are you?” Pothead: “No officer, its ‘Hi, how are you’”
Tefft BC et al. Prevalence of Marijuana Involvement in Fatal Crashes: Washington, 2010 – 2014. May 2016. Washington, DC: AAA Foundation for Traffic Safety. [free full text]
This is a report by the AAA Foundation for traffic safety. It retrospectively looked at a database from the Washington State Traffic Safety Commission. In Washington State, as many will know, marijuana became legal in December of 2012. This study looked at all motor vehicle collisions that resulted in death and the proportion who had THC (delta-9-tetrahydrocannabinol) in their blood on autopsy. They compared collisions in the 2 years before the new law to the 2 years after the law. Out of the total of 3031 fatal MVCs over 4 years, 303 (10%) involved drivers testing positive for THC. The percentage rose from 8.9% in 2013 (before the law) to 17% in 2014. Of the individuals with positive tests for THC, 39% also had alcohol on board, 16% had other drugs, and 10% had alcohol and other drugs (leaving 34% with only THC detected). There are some problems with this data, the biggest probably being that we don’t know what levels of THC correlate with being impaired. THC wasn’t measured in every case, and sometimes measurement was delayed. Also, the total number of fatalities didn’t increase, just the number with THC on board, so the marijuana could be a bystander rather than a cause of the collisions. However, the issue of impaired driving, and our lack of science to guide us, remains a huge issue as this popular legalization movement continues forward.
Bottom line: Don’t smoke and drive
Would you be surprised if this patient died?
George N, Phillips E, Zaurova M, Song C, Lamba S, Grudzen C. Palliative Care Screening and Assessment in the Emergency Department: A Systematic Review. Journal of pain and symptom management. 51(1):108-19.e2. 2016. PMID: 26335763
I really hate adding work for our overly taxed triage nurses, who end up doing a lot of our screening. However, we are awful at recognizing patients with palliative care needs. This is a review, and I don’t think it is strong enough to completely overhaul triage systems, so I won’t go into the details, but they do conclude that palliative care screening is feasible. My favorite screening question is “would you be surprised if this patient died during this visit or in the coming month?” If no, they should probably have palliative care involved. Of course, the harder part of this equation is actually having adequate palliative resources for all the patients who need them.
Bottom line: Emergency medicine is all about dying patients. Palliative care should be an essential part of our mindset. You can have a much bigger impact by starting palliative care than you ever will handing out antibiotics for sore throats.
Epinephrine is safe in fingers – is that old news by now?
Ilicki J. Safety of Epinephrine in Digital Nerve Blocks: A Literature Review. The Journal of emergency medicine. 49(5):799-809. 2015. PMID: 26254284
This is a systematic review looking at the use of epinephrine in digital blocks. It should be stated up front that the conclusions are only good as the original studies – and they aren’t great. In a total of 30 studies, they identify 2797 blocks performed with epinephrine without any complications. They conclude that epinephrine is safe to use in digital nerve blocks in healthy patients. (Although, to be fair, it probably doesn’t help most of the time.)
Bottom line: This is probably another classic myth, although the data isn’t actually strong enough to definitely conclude safety.
CT first for the scaphoid?
Yin ZG, Zhang JB, Gong KT. Cost-Effectiveness of Diagnostic Strategies for Suspected Scaphoid Fractures. Journal of orthopaedic trauma. 29(8):e245-52. 2015. PMID: 25756914
I hate cost-effectiveness studies. The results hinge on a huge number of assumptions that really can’t be confirmed. However, sometimes they provide some insight that can be interesting. In this study they compared the cost of working up scaphoid fractures using a number of different strategies, such as immediate CT, immediate MRI, MRI on day 3, bone scan on day 3, and x-ray at 2 weeks. Maybe counterintuitively, the immediate CT was the most cost effective approach followed by the immediate MRI. I am not sure that you can take these results to the bank, but it is a good reminder that there a number of costs that we often forget about. Although immediate CT seems expensive, to come to a follow-up visit the patient must miss work, pay for parking, and see another physician. Additionally, being in an unnecessary cast for 2 weeks could result in 2 weeks unnecessarily off work. It’s not time to change yet, but I wouldn’t be surprised if immediate definitive testing for the scaphoid became the standard in the future.
Bottom line: CT first for scaphoid fracture might actually be cheaper than standard practice.
Compassion and the good Samaritan study
Darley JM, Batson CD. “From Jerusalem to Jericho”: A study of situational and dispositional variables in helping behavior. Journal of Personality and Social Psychology. 27(1):100-108. 1973. [article]
This is a classic study from the psychology literature. It is a study of seminary students. Half were told they had to give a talk on the parable of the good Samaritan. The other half were told they had to give a talk on routine seminary jobs. Some students were told they were late for the talk and had to hurry, whereas others were told that they were on time. There was a plant on the way to the lecture hall: a man slumped in a doorway who moaned twice as the students walked by. Whether or not they had been thinking about the parable of the good Samaritan made no difference in whether or not they stopped. The only thing that influenced their decision to help was how hurried they felt. I see a clear connection to emergency medicine. Our job requires an incredible amount of compassion. However, simply thinking about compassion doesn’t seem to help. If we are rushed, we are less likely to be compassionate, whereas if we have time, we will use it. Unfortunately – how many people feel like they have a lot of free time in the emergency department? Managing our departments so physicians are not constantly run off their feet is probably a really good idea. (Of course, that is quite a reach from this research scientifically speaking – but it makes a lot of sense to me.)
Bottom line: Ensuring that physicians aren’t rushed might be crucial in increasing our ability to be compassionate on the job.
Cheesy Joke of the Month
I was sitting in a bar the other night when the waitress yelled out “does anyone know CPR?!”
I yelled back, “yeah, and I know the rest of the alphabet too.”
Everyone laughed. Well, except one guy.
#FOAMed of the month
How is it that one develops mastery? Cliff Reid provides some insight into deliberate practice in the context of getting his ass kicked.