Amoxicillin is the antibiotic of choice in pediatric pneumonia
Williams DJ et al. Narrow vs broad-spectrum antimicrobial therapy for children hospitalized with pneumonia. Pediatrics. 2013 Nov;132(5):e1141-8. PMID: 24167170
This was a retrospective record review of 15,564 admitted but not critically ill pediatric patients with community acquired pneumonia. They used propensity scoring, so the results could mean anything, but kids getting amoxicillin had the same outcomes as those with broad spectrum antibiotics such as cefotaxime or ceftriaxone. In fact, IDSA and peds infectious disease society both recommend narrow spectrum antibiotics, which is contrasted to the 90% of children in this study that were given broad spectrum.
Bottom line: Amoxicillin is probably best in pediatric pneumonia.
Hans and Franz want to pump you up (steroids for pediatric asthma)
Keeney GE et al. Dexamethasone for acute asthma exacerbations in children: a meta-analysis. Pediatrics. 2014;133(3)493-9. PMID: 24515516
A meta-analysis of 6 RCTs of prednisone versus dexamethasone in children with acute asthma exacerbations. There was no difference in relapse at 5 or 30 days. The dexamethasone group was less likely to vomit, both at home and in the ED. (Some studies used 2 doses of dex, some only used 1 versus generally 5 days of prednisone.)
Bottom line: Fewer doses and less vomiting, I am sold on dexamethasone. (My wife adds: “Well Duh! Pediapred tastes like s***. Dex is less volume and way easier to take.”)
The ugly stepchild of papers 1 and 2? Steroids for pneumonia
Blum, CA et al. 2015. Adjunct prednisone therapy for patients with community-acquired pneumonia: a multicentre, double-blind, randomised, placebo-controlled trial. Lancet (January 16). PMID: 25608756
I don’t buy what they are selling here, but I have already heard about this paper from at least 10 different sources, so you will likely hear about it as well. This is a large, multi-center, double blind RCT of 781 community acquired pneumonia patients, randomized to either get or not get prednisone 50mg PO daily for 1 week. It was a positive study, in that the primary outcome “time to clinical stability”, or ‘normal vital signs’, was 3 days in the prednisone group and 4.4 days with placebo. However, as important as vital signs are, are they really a patient oriented outcome? Has a patient ever said, I know I have this pneumonia, but what I really want is for my heart rate to be 95 instead of 105? Side effects: prednisone obviously caused hyperglycemia, but also (non statistically) doubled pneumonia associated complications. Previous studies showed higher recurrence rates with steroids.
Bottom line: Of course steroids make the numbers look better, but we are probably treating the doctor and not the patient here. Not for me.
Bottom line #2: If you are going to design a study, measure outcomes that matter.
Why do we use cervical collars?
Ala’a O et al. 2015. Should suspected cervical spinal cord injury be immobilised?: A systematic review. Injury Journal. (In press). PMID: 25624270
Like many of the things we do, this practice was started based on expert opinion in the pre-EBM era. There are a large number of cadaver and volunteer studies that show that C-collars really don’t prevent movement of the c-spine. What is the clinical evidence? There are a grand total of 8 observational studies ever done. In penetrating trauma, C-collar application was associated with an increase in mortality (OR 8.8), increase scene time, and concealment of neck injuries. In blunt trauma, one study showed that immobilization was associated with worse neurological outcomes. This is balanced by no evidence of benefit. They conclude “there is a clear need for large prospective studies to determine the clinical benefit of prehospital spinal immobilsation.”
Bottom line: I can’t imagine anyone changing their practice, but this does not speak very well to the benefits of cervical spine collars
Where are you drilling? Arm might be better than leg, or go straight towards the heart
Pasely J et al. 2015. Intraosseous infusion rates under high pressure: A cadaveric comparison of anatomic sites. Journal of Trauma and Acute Care Surgery 78(2)295-9. PMID: 25757113
Its a cadaver study, so take that as you will – but I am often drilling into dead people in code situations anyhow, so there might be some external validity here. They tried to infuse saline using a pressure bag, and the rates they could get were: 94ml/min in the sternum, 57ml/min in the humerus, and 30 ml/min in the tibia.
Bottom line: Humerus seems twice as fast as the tibia, so maybe that should be our go to spot? I probably wouldn’t suggest drilling sharp things into the sternum, but some people seem to think it’s OK.
Speaking of IOs – they are fine for RSI
Barnard EBG et al. 2014. Rapid sequence induction of anaesthesia via the intraosseous route: a prospective observational study. Emerg Med J (electronic ahead of print). PMID: 24963149
OK, also not really definitive by any means. A prospective observational study, with no controls, in a military setting. 34 patients had their RSI drugs pushed through an IO, first pass success in all but 1 (97%) and that patient was intubated on the second attempt. Although no control, 97% compares well with historical controls.
Bottom line: Go ahead and give RSI drugs through an IO if that is what you have
First RCT of massive transfusion protocol
PROPPR Holcomb et al. Transfusion of Plasma, Platelets, and Red Blood Cells in a 1:1:1 vs a 1:1:2 Ratio and Mortality in Patients With Severe Trauma. The PROPPR Randomized Clinical Trial. JAMA. 2015; 313(5)471-82. PMID: 25647203
After a bunch of theoretical stuff and some observational trials, this was the first ever RCT comparing different ratios of PRBCs, FFP, and platelets in a massive transfusion protocol. They compared 1:1:1 PRBCs, FFP and platelets to 2 units of PRBCs for each 1 unit of FFP and platelet equivalent. This was a negative trial, in that there was no difference in mortality between the two groups. However, some people have argued that their goal of a 10% reduction in mortality was too high, that the non-significant trends (including a 4.3% absolutely mortality reduction) favoured the 1:1:1 group, and secondary bleeding end points also favoured the 1:1:1 group. (This study design makes the inherent assumption that some transfusion ratio is a good thing, in that they did not include a usual care arm. While this has been the trendy thing of late, it is entirely based on flawed observational studies.)
Bottom line: This study will be used to support whatever pre-existing beliefs you had on the subject.
The new AAP bronchiolitis guidelines are very nihilistic (maybe realistic?)
Ralston SL et al. 2014. Clinical practice guideline: the diagnosis, management, and prevention of bronchiolitis. Pediatrics 134(5)e1474-502. PMID 25349312
Do NOT give ventolin
Do NOT give epinephrine
Do NOT give hypertonic saline (in the ED)
Do NOT give corticosteroids
Diagnosis on Hx/Px, no routine chest xrays
While these guidelines are very evidence based, my EBM self is fighting with my practical self. If there are no treatments, peds is going to have to see 30 kids a day in the ED. Should we just set aside a room for them?
Bottom line: The AAP says don’t do anything for bronchiolitic kids
Two for the price of one: pediatric head injuries aren’t cured by CT
Lee LK et al. (PECARN). Isolated loss of consciousness in children with minor blunt head trauma. JAMA Pediatrics 2014; 168(9)837-43. PMID: 25003654
This is a secondary analysis of the PECARN head injury algorithm. Although overall your chance of clinically important head injury was 2.5% with LOC and only 0.5% without, if you only had LOC and no other PECARN risk factors, your risk of a clinically important injury was the back to baseline at 0.5%.
Bottom line: Loss of consciousness, in the absence of other worrisome findings, has a low risk of clinically important injury and CT scan is unnecessary. (Look at the whole patient, not just one aspect of the history or physical.)
Dayan PS et al. (PECARN). Association of traumatic brain injuries with vomiting in children with blunt head trauma. Annals of Emergency Medicine 2014;63(6)657-65. PMID: 24559605
Another secondary analysis of the PECARN head injury algorithm. Vomiting, without any other PECARN risk factors, had an overall incidence of clinically important injury of 0.2%
Bottom line: Vomiting, in the absence of other worrisome findings, has a low risk of clinically important injury and CT scan is unnecessary. (Look at the whole patient, not just one aspect of the history or physical.)
Start sending those stroke patients to the cath lab?
After multiple negative trials in the past, we get 3 new trials on endovascular treatment of stroke. (Given that we aren’t a stroke center and this isn’t going to be a decision you will make in the ED, it is probably best to just skip to the next section. But they will be talked about at cocktail parties.)
MR CLEAN Berkhemer OA et al. A randomized trial of intraarterial treatment for acute ischemic stroke. N Engl J Med. 2015;372:(1)11-20. PMID: 25517348
RCT comparing intra-arterial treatment versus usual care in stroke patients. Good neurological outcome (MRS 0-2 at 90 days) in intra-arterial group was 32% versus only 19% in the usual care group. (These are both way worse outcomes than other stroke trials, like NINDS)
EXTEND-IA Campbell BC et al. Endovascular Therapy for Ischemic Stroke with Perfusion-Imaging Selection. N Engl J Med. 2015. (Ahead of print) PMID: 25671797
RCT (phase II trial) of patients getting TPA within 4.5 hours with a middle cerebral or internal carotid clot AND evidence of salvageable brain tissue plus or minus endovascular therapy. Was stopped early after only 70 patients (they had to screen over 7,000 patients at 10 hospitals over 2 years to find these 70 patients – so they are highly selected to say the least). There were multiple primary outcomes (bad) but importantly if you got treated 80% had good neurological improvement at 3 days, versus only 37% of those without the endovascular treatment.
ESCAPE Goyal M et al. Randomized Assessment of Rapid Endovascular Treatment of Ischemic Stroke. N Engl J Med. 2015. (Ahead of print) PMID: 25671798
RCT of patients up to 12 hours with proximal anterior circulation occlusions and evidence of good collateral flow plus or minus endovascular therapy. Also stopped early, with a total of 316 patients (wanted 500 originally). They also only managed to recruit about 1 patient a month at each of the 22 hospitals involved – so also very highly selected patients. Functional independence (MRS 0-2) at 90 days was 53% in the endovascular arm and 29% in the usual care arm.
Overall bottom line: The benefit described in these trials is impressive. They are small and all have some flaws (stopping them early probably exaggerates the benefit), but I think it is likely they represent a true benefit. However, the number of eligible patients was tiny. Maybe they have finally found the subset of stroke patients that will benefit from revascularization – like the STEMI patient in a sea of chest pains.
Dr. Oz Sucks
Korownyk C et al. Televised medical talk shows–what they recommend and the evidence to support their recommendations: a prospective observational study. BMJ 2014;349:g7346. PMID: 25520234
OK, this isn’t really all that valuable or surprising, because anyone that has ever turned on a TV realizes that Dr. Oz rarely has anything credible to say, and seems to be a lot more interested in selling snake oil than actually helping patients. But in case any one was wondering, these authors prospectively evaluated the claims made on Dr. Oz and The Doctors, and even if a single case report was counted as “evidence” only 50% of the claims made on the shows had any evidence based backing, and a full 15% were completely contradictory to available evidence.
Bottom line: Don’t get your medical advice from a TV shill
Let’s review an older one: TTM, putting dead people on ice
Nielsen N et al. Targeted temperature management at 33°C versus 36°C after cardiac arrest. N Engl J Med. 2013 369(23):2197-206. PMID: 24237006
An ‘older’ paper that I am sure everyone has heard about, but it is good to include at least one practice changing quality study every month. After 2 small, low quality studies were published in 2002 (well before I started medical school in case you were wondering), the medical world went nuts for therapeutic hypothermia. But when I started in medicine, there were still some intelligent people (like Jerry Hoffman) who tried to remind us these were small studies, with inherent biases, and that a corner stone of science is replication. (There is a lesson here for so many other topics – but I don’t think I have the balls to mention NINDS and tPA.)
So this was a large, randomized control trial (not blinded) where 950 patients with ROSC after out of hospital cardiac arrest were either brought to 33 or 36 degrees Celsius. There was no difference in outcome.
The comments about this paper have been all over the map. The favorite statement by a lot of very smart people seems to be “this confirms that we desperately need to avoid fever, but 36 degrees is probably good enough.” I would point out, this study says nothing about avoiding fever. In fact, I don’t know of any study that compared fever or no fever post cardiac arrest. So people are either expressing their left over love of hypothermia, or is basing it on animal models, which are – well animal models.
Another approach would be to ask if we have any reason to believe this would work (the beginning of Bayesian reasoning). There were some animal models that support hypothermia, but probably more important is that hypothermia has been tested in humans for a number of conditions other than cardiac arrest – and it doesn’t seem to work.
Bottom line: There is no benefit from hypothermia post cardiac arrest. No one knows much about fever, but many people will talk about it a lot.
Bonus section: This Penn and Teller vaccination video should play continusouly in the waiting room
Cheesy Joke of the Month
It was a cold February so:
What is the difference between snowmen and snowwomen?
Justin Morgenstern. Articles of the month (February 2015), First10EM, 2015. Available at: