Articles of the Month (December 2016)

Welcome to the year-end edition of the Articles of the Month (released well into the new year because of the craziness of emergency department holiday schedules). The podcast version of this post with Casey Parker is available through the BroomeDocs podcast.

Continue reading “Articles of the Month (December 2016)”

Articles of the Month (March 2016)

My monthly summaries of the medical literature

Every month I select the best medical articles I have read and provide brief summaries and critical appraisals. Here are this month’s articles:

The paper you are most likely to hear about this month: antibiotics and abscesses

Talan DA, Mower WR, Krishnadasan A. Trimethoprim-Sulfamethoxazole versus Placebo for Uncomplicated Skin Abscess. The New England journal of medicine. 374(9):823-32. 2016. PMID: 26962903

Until now, the data on abscess management has been pretty clear: all you need is cold hard steel. No packing, and definitely no antibiotics. Has management just become much more confusing? This is a large, multi-center RCT comparing trimethoprim-sulfamethoxazole (320mg/1600mg MID for 1 week) to placebo in 1247 adult patients with acute abscess greater than 2cm in diameter. For the primary outcome of clinical cure at 7 days, the antibiotics group was better (80.5% versus 73.6%; absolute difference 6.9% 95%CI 2.1-11.7%; NNT = 14). There were also decreases in several secondary outcomes, such as new skin infections at other sites. However, there was an increase in GI side effects by about 7% (42.7% vs 36.1%). A cure rate of only 75% is really low and doesn’t represent the patients I see. This is probably because these are not simple abscess, with a median cellulitis area of 6.5×5.0cm, 20% of the cohort with a cellulitis area greater than 75cm2, and many patients “met other guideline criteria for antibiotics treatment”. (You can read some other opinions on REBEL EM, EM Nerd, and EM Literature of Note.)

Bottom line: This isn’t a game changing paper. It tells you to keep using antibiotics in the patients you are already using them in – complex abscesses with cellulitis – and doesn’t tell us a lot about the average abscess.

How ready are you for a mass casualty event?

Bhalla MC, Frey J, Rider C, Nord M, Hegerhorst M. Simple Triage Algorithm and Rapid Treatment and Sort, Assess, Lifesaving, Interventions, Treatment, and Transportation mass casualty triage methods for sensitivity, specificity, and predictive values. The American journal of emergency medicine. 33(11):1687-91. 2015. PMID: 26349777

I found this paper fascinating. I won’t get into detail about the performance of the scores, because the data is retrospective, and there is too much information that these scores use that would not be well recorded. However, I think this is a great study to read. I had never been exposed to a mass casualty triage system before, nor do I think I have been adequately trained in this aspect of emergency medicine. The algorithms are interesting. It’s worth a read.

Bottom line: Are you for a mass casualty event? This article might help

How do you tell if a patient needs more pain medication? Ask them

Chang AK, Bijur PE, Holden L, Gallagher EJ. Efficacy of an Acute Pain Titration Protocol Driven by Patient Response to a Simple Query: Do You Want More Pain Medication? Annals of emergency medicine. 2015. PMID: 26074387

I am strongly biased to like this study, because I was always thaught this is the way that pain medicine should be given (by Jerry Hoffman, I think). This is a prospective cohort of 215 adult patients presenting to the ED for acute painful conditions that the attending physician thought required an opioid. All patients received hydromorphone 1mg IV, followed by up to 3 more 1mg IV doses every 30 minutes driven entirely by their response to the question “Do you want more pain medication?” There were delays in administration of medication, so the max was actually 4mg over 4 hours. 205 of the 207 patients (99% 95%CI 97-100%) achieved pain control on 1 or more occasions during the study; 97% were either satisfied or very satisfied with their pain treatment. About 20% of patients wanted more pain meds at each interval. What can we learn from this study? The design of the study allows us to conclude that this method will leave most people satisfied with their pain control, and that almost everyone will reach a point where they don’t want any more analgesia. However, with no comparison, we have no idea if this technique is any better or worse than other methods. Personally, I am interested in how this would compare to patient controlled analgesia – which would require a larger upfront investment, but I think would be more tolerable for nursing workloads.

Bottom line: Asking patients is a reasonable method to determine if they need more analgesia

2 is not be better than 1 when in comes to needles

Martin SP, Chu KH, Mahmoud I, Greenslade JH, Brown AF. Double-dorsal versus single-volar digital subcutaneous anaesthetic injection for finger injuries in the emergency department: A randomised controlled trial. Emergency medicine Australasia : EMA. 2016. PMID: 26991958

Injections in the palm always seemed painful to me, so I always stuck with the double dorsal injection technique. I don’t remember why I changed, but my success rate is much better with the single palmar injection, so I’ve never looked back. The study: 86 adult patients in an RCT comparing a double-dorsal to a single-palmar injection technique for digital nerve block. There was no difference in the pain of injection between the two techniques (almost 4/10). The techniques were equally successful (65% success with double-dorsal and 72% with single-palmar). Really, none of those numbers are great.

Bottom line: Stick with the bloc you are used to – and maybe add some bicarb to get the pain on injection down?

Dumb and dumberer

Maltese F, Adda M, Bablon A. Night shift decreases cognitive performance of ICU physicians. Intensive care medicine. 42(3):393-400. 2016. PMID: 26556616

This is a prospective, randomized, cross-over study of 51 ICU doctors (27 residents, 21 attendings) who were randomized to either work a night shift or rest at home (and then were crossed over to the opposite group). Between 10am and noon the next day they went through a series of psychological tests. Not surprisingly, working memory, information process speed, and perceptual reasoning were all worse after a night shift. Cognitive flexibility was not statistically different. The clear issue with the study is we have no idea how these psychological tests translate into patient care, or whether the measured differences are actually clinically important differences.

Bottom line: Night shifts are hard. It’s hard to make good decisions at 4am. (One of many reasons I like the idea of casino shifts).

How safe is that treatment really?

Saini P, Loke YK, Gamble C, Altman DG, Williamson PR, Kirkham JJ. Selective reporting bias of harm outcomes within studies: findings from a cohort of systematic reviews. BMJ (Clinical research ed.). 349:g6501. 2014. PMID: 25416499 [free full text]

These authors attempt to examine the accuracy of reporting of harms in clinical trials. They identified trials that had been included in systematic reviews, and then evaluated each trial for how they reported harms (ie, harms measured and reported, harms not measured, harms measured but only partially reported, harms not even mentioned…). When looking at all Cochrane reviews, they found that the studies only partially reported or didn’t report harms at all 76% of the time. In a group a systematic review designed specifically to look at adverse events, 47% of studies still did not report or only partially reported a single primary harm outcome. This tendency of the literature has been discussed before. We tend to minimize our discussion of harms, which obviously skews our conclusions when looking at the entirety of the literature.

Bottom line: For every medicine you use, remember that the harms are probably greater than those reported in clinical trials

Clinical correlation required

Mark DG, Sonne DC, Jun P. False negative interpretations of cranial computed tomography in aneurysmal subarachnoid hemorrhage. Academic emergency medicine : official journal of the Society for Academic Emergency Medicine. 2016. PMID: 26918885

This is a chart review that identified 452 patients diagnosed with subarachnoid hemorrhage, and looked at the subset of 18 patients (4%) who were diagnosed by lumbar puncture after a normal CT. The supposedly normal CTs were then reviewed by 2 speciality neuro-radiologists, and in 9 of the 18 (50%) the neuro-radiologists thought there was evidence of bleed on the scan initially reported as normal. The false negative rate was 71% (5 of 7) for the scans done within 6 hours of headache. Of course, re-reading studies without the time pressures of a normal shift, especially when LPs have already shown blood (although these radiologists were blinded) might be easier.

Bottom line: Radiology isn’t perfect. Not all radiologists are created equal. This might still be the fatal flaw in the famous Perry study

But doc, it hurts A LOT

Body R, Lewis PS, Carley S, Burrows G, Haves B, Cook G. Chest pain: if it hurts a lot, is heart attack more likely? European journal of emergency medicine : official journal of the European Society for Emergency Medicine. 23(2):89-94. 2016. PMID: 25340995

Does the pain score correlate with the chance of MI? This is a secondary analysis of a data set collected for a prospective cohort study that included emergency department patients with suspected cardiac chest pain. They looked at the pain scores of the patients with a final diagnosis of MI, as compared with those who ruled out. Although there was a statistical difference, with the average pain in the MI group being 8/10 (interquartile range 5-8) and the non-MI group being 7/10 (IQR 6-8) (p=0.03), those numbers obviously don’t help clinically. This is reinforced by their analysis that showed the area under the receiver operating curve was 0.58, so essentially a coin flip. The amount of pain might have influenced the original physicians in terms of who was included in the dataset, which would skew these numbers.

Bottom line: The intensity of pain does not seem to help diagnostically in ACS.

Not the worry, that blurry vision and headache should be gone in … 90 days?

Kriz PK, Stein C, Kent J. Physical Maturity and Concussion Symptom Duration among Adolescent Ice Hockey Players. The Journal of pediatrics. 2016. PMID: 26781190

How long do pediatric concussion symptoms last? This is a prospective cohort of 145 patients aged 13-18 years who were referred to a sports medicine clinic. The mean symptom duration was 45 days (though with wide confidence intervals of +/- 49 days). About half (48%) of patients had symptoms for more than 28 days, and 13% had symptoms beyond 90 days. So perhaps concussion symptoms last longer than we usually counsel, but I worry about a significant selection bias here, as we don’t refer most kids with concussion to sport medicine clinics, and this is only 145 children from 3 clinics over 2.5 years. It probably represents the worst case scenario.

Bottom line: Some children will have prolonged concussion symptoms. Counselling and follow up instructions should keep this in mind

I have to say, the heart is what won me over when it comes to POCUS

Martindale JL, Wakai A, Collins SP. Diagnosing Acute Heart Failure in the Emergency Department: A Systematic Review and Meta-analysis. Academic emergency medicine : official journal of the Society for Academic Emergency Medicine. 2015. PMID: 26910112

This is a systematic review looking the diagnosis of congestive heart failure in the emergency department. Probably because CHF actually encompasses a few different underlying pathologies, their major finding was that no factors were good enough to single handedly rule in or rule out CHF. Some likelihood ratios: Audible S3 +LR 4.0, CXR signs of edema +LR 4.8, B lines on bedside US +LR 7.4, no B-lines -LR 0.16, and reduced ejection fraction on bedside echo +LR 4.1. (Remember you want a positive likelihood ratio of 10 or more to rule in, and a negative likelihood ratio of 0.1 or less to rule out.)

Bottom line: Ultrasound may be our best tool for diagnosing CHF, but no findings can be used in isolation to rule in or rule out the disease.

There are now bottles of water labeled “gluten free”

Zanini B, Baschè R, Ferraresi A. Randomised clinical study: gluten challenge induces symptom recurrence in only a minority of patients who meet clinical criteria for non-coeliac gluten sensitivity. Alimentary pharmacology & therapeutics. 42(8):968-76. 2015. PMID: 26310131

I get asked a lot about gluten, as I am sure many people do, not just by emergency medicine patients, but also friends and families. In this study they took 35 patients who had tested negative for celiac disease but self-identified as being gluten intolerant and in a double blind, cross-over design they exposed the patients to either gluten free or normal flour. 12 (49%) of the patients thought the gluten-free flour contained gluten – what you might guess if the guess was pure chance. I think the evidence is pretty convincing that gluten is not the issue for most people without celiac disease. These patients definitely have symptoms, but there are almost certainly other etiologies than gluten. I worry that the focus on gluten pushed by some ‘experts’ is doing patients a disservice.

Bottom line: In people without celiac disease, symptoms are not consistently reproduced by gluten.

The case of the killer iPhone

Tri JL, Severson RP, Hyberger LK, Hayes DL. Use of cellular telephones in the hospital environment. Mayo Clinic proceedings. 82(3):282-5. 2007. PMID: 17352363

These authors brought 2 telephones into different patient rooms and made phone calls while observing various medical devices. In total they did 300 tests of 192 difference devices in 75 patient rooms, and they were unable to document a single case of the cellular telephone interfering with medical equipment.

Bottom line: The ban on cel phones may be based on a myth?


Lawrentschuk N, Bolton DM. Mobile phone interference with medical equipment and its clinical relevance: a systematic review. The Medical journal of Australia. 181(3):145-9. 2004. PMID: 15287832 [free full text]

This is a systematic review looking at the question of mobile phone interference with medical equipment. They identified 8 studies that tested a total of 936 devices. (Studies ranged from 1994-2002, so technology may have changed since.) They found that interference did occur in as many as 6% of tests. However, essentially all of this interference occurred when the phone was within 1 meter of the device – so you probably have to be trying to cause interference, like they were in these studies, rather than just using phones normally. Also, the results of the interference were not recorded, so it’s difficult to know if any of it was clinically relevant. (Some brief interference on an ECG monitor is irrelevant, but I would care about a pacemaker that stopped pacing.)

Bottom line: This is a little more complex than the last paper indicated, but it appears phones are safe as long as they are more than a meter from medical equipment. (Although that might be hard in some of the cramped resuscitation rooms I have worked in.)

NSAIDs for 11/10 pain

Pathan SA, Mitra B, Straney LD. Delivering safe and effective analgesia for management of renal colic in the emergency department: a double-blind, multigroup, randomised controlled trial. Lancet (London, England). 2016. PMID: 26993881

This is a double blind RCT of 1644 adult patients with renal colic (1316 confirmed on CT) comparing morphine (0.1mg/kg IV) to diclofenac (75mg IM) to acetaminophen (1 gram IV). For a primary outcome of a 50% reduction in pain at 30 minutes, diclofenac was more effective than either morphine or acetaminophen, which weren’t different from each other (OR 1·35, 95% CI 1·05-1·73, p=0·0187). This means that 68% of the diclofenac group had a 50% reduction in pain, as compared to 61% with morphine and 66% with acetaminophen – not a huge absolute difference. One interesting number is that only 12% of the IM diclofenac group needed any rescue medication, so it might be possible to manage renal colic without ever starting an IV. Adverse events were statistically higher in the morphine group, but really quite low (1-3%) in all groups. Personally, I like a multimodal pain approach, and will probably continue to combine NSAIDS and opioids.

Bottom line: It might be true that NSAIDs are slightly more effective in renal colic

#FOAMed of the Month

This is a really short post on the Nurse Path, but I love it because it is a simple yet brilliant method for improving communication and patient safety. The key is that for medication checks, rather than reading out the dose and asking the person confirm ‘yes or no’, which could result in confirmation bias or error, you simply ask “what is this?” That forces the other person to slow down and actually read the medication out loud. I imagine this technique could also be used in another of other situations as well.



Cheesy Joke of the Month

Two orthopedic surgeons are on opposite sides of a lake.

One surgeon yells to the other, “How do you get to the other side?”

The other responds, “You are on the other side!”

Articles of the month (February 2016)

There are new sepsis definitions! Hurrah?

Singer M, Deutschman CS, Seymour CW. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA. 315(8):801-10. 2016. PMID: 26903338 [free full text]

There are new sepsis guidelines. I guess that warrants headline news, and there has been a lot of excitement on the medical internet. However, they are really just the opinions of 19 experts, aren’t backed by any quality prospective data, and probably shouldn’t change your management. If you want to read more, I wrote a full post on the topic: Sepsis 3.0 – No thank you

Bottom line: Talk about qSOFA if you want to sound in the know, but clinically I would ignore this paper

Procedural sedation consent: “Don’t worry, it’s super safe… it’s the Michael Jackson drug.”

Bellolio MF, Gilani WI, Barrionuevo P. Incidence of Adverse Events in Adults Undergoing Procedural Sedation in the Emergency Department: A Systematic Review and Meta-analysis. Academic emergency medicine : official journal of the Society for Academic Emergency Medicine. 23(2):119-34. 2016. PMID: 26801209

What exactly are the risks of procedural sedation? I know them qualitatively, but when having an informed choice conversation, are you able to quote the actual incidence? I know I couldn’t. This is a systematic review and meta-analysis to determine the incidence of adverse events in ED procedural sedation (limited to after 2004). They found 55 articles that covered 9652 procedural sedations. The most common adverse events: hypoxia (40/1000 but only 23/1000 were <90%), vomiting (16/1000), hypotension (15/1000), and apnea (12/1000). The serious adverse events: laryngospasm (4/1000), intubation (1.6/1000), aspiration (1.2/1000). If you are interested, they do break some of these numbers down based on what agent was used. There was a fair amount of heterogeneity in the definitions used in the original studies. Also pediatrics was excluded.

Bottom line: Procedural sedation is safe, but we should have a sense of these numbers for adverse events.

Still not using topical anesthetics for corneal abrasions? Could topical NSAIDs be a better choice?

Calder LA, Balasubramanian S, Fergusson D. Topical nonsteroidal anti-inflammatory drugs for corneal abrasions: meta-analysis of randomized trials. Academic emergency medicine : official journal of the Society for Academic Emergency Medicine. 12(5):467-73. 2005. PMID: 15860701 [free full text]

Most people have heard me rant about the myth that topical anesthetics are harmful in corneal abrasions. (If you haven’t, watch for an upcoming episode of EMCases Journal Jam, or come to the North York General Emergency Medicine Update this year.) However, an essential part of informed choice is reviewing the alternatives. How do topical NSAIDs perform in managing the pain of corneal abrasions? (Hat tip to Nadia Awad @Nadia_EMPharmD for sending me this paper.) This is a systematic review and meta-analysis that identified 11 RCTs (they don’t report the total sample size, but they were all relatively small studies). I find this paper a little hard to follow, because they report 5 high quality studies to be included in the meta-analysis, but then include only 3 in the forrest plot. Looking at just these 3 trials (n=459), topical NSAIDs did decrease pain, with a weighted mean difference of -1.30 (95%CI -1.56 to -1.03) on a 10 point pain scale. There are a few issues with this data. First: it’s hard to interpret a weighted mean difference, but the minimum change on a 10 point pain score generally considered to be clinically important is 1.4. Second: there is a lot of data that could not be included because of the way the original trials were reported. Third: although a formal funnel plot couldn’t be done, the authors admit a possibility of publications bias. Fourth: There is not enough data on safety, but there was at least one recurrent corneal erosion in the NSAID group. Fifth: The funding source of the original trials was not discussed, but it might be important considering that not a single one of the trials had allocation concealment. Finally: the comparison groups were varied, but often just placebo. It might be better to compare to the less expensive oral NSAIDs (or topical anesthetics.)

Bottom line: Topical NSAIDs may decrease pain from corneal abrasions, but I don’t think this data is enough to support using them over other agents (especially considering their cost.)

Xanthrochromia AKA hey Bob, does this look kinda yellow to you?

Chu K, Hann A, Greenslade J, Williams J, Brown A. Spectrophotometry or visual inspection to most reliably detect xanthochromia in subarachnoid hemorrhage: systematic review. Annals of emergency medicine. 64(3):256-264.e5. 2014. PMID: 24635988

This is a systematic review looking at studies (English only) that included patients presenting with a headache who had LPs where the CSF was sent for xanthrochromia. The gold standard for SAH was either angiography or follow up (not perfect). The studies were also highly heterogenous. Not surprisingly, visual inspection, AKA “hey Bob, does this look kinda yellow to you”, was not perfect, with a sensitivity of 84%, specificity of 96%, positive LR of 14.1 and negative LR of 0.35. However, the fancy spectrophotometry was not any better, with a sensitivity of 87%, specificity of 86%, positive LR of 6.6 and negative LR of 0.29. The included studies are not of high enough quality to be sure about any of those numbers. I just don’t understand how we don’t have something better yet – obviously some chemical is turning the fluid yellow – could the makers of super-ultra-sensitive troponins not just create a test that detects whatever this compound is?

Bottom line: Neither method of detecting xanthochromia is perfect, which adds another layer of complexity to the question of who we should be LPing after CT

Foley free pee?

Herreros Fernández ML, González Merino N, Tagarro García A. A new technique for fast and safe collection of urine in newborns. Archives of disease in childhood. 98(1):27-9. 2013. PMID: 23172785

Here is a contribution from Dr. Kate Bingham. You probably know how I feel about getting urines in pediatric patients. (If you don’t, you can read this.) However, for newborns, a urine culture is going to get done. This paper describes a technique to get the urine without a foley. Basically, feed kid, wait 25 min, clean genitals, hold baby under armpits (standing position), tap suprapubic area at 100/min for 30 seconds, then massage low back for 30 seconds. Repeat until pee is produced, and make sure you catch it in specimen bottle. Does it work? Of the 80 patients they tried this on (no comparison group), they were successful in 69 (86%). Median time to sample collection was 45 seconds. My only concern is if I miss the urine and I have to start all over again (maybe after antibiotics). This is interesting, but I so rarely get newborn urines, I will probably stick with a Foley for now.

Bottom line: You can make children pee using this technique. Not sure where to fit that into practice.

I never get tired of talking about nerve blocks

Dickman E, Pushkar I, Likourezos A. Ultrasound-guided nerve blocks for intracapsular and extracapsular hip fractures. The American journal of emergency medicine. 2015. PMID: 26809928

One rebuttal I have often encountered when talking about nerve blocks for hip fractures is that the block is less likely to work in certain fracture patterns. This is a secondary analysis of data from a previously conducted prospective RCT looking at 77 patients and comparing the effectiveness of ultrasound guided femoral nerve block in intracapsular versus extracapsular hip fractures. They were the same, and both were good (pain scores from 6.5/10 just under 4/10 at 2 hours).

Bottom line: I will keep using nerve blocks for all hip fractures. I’m not too worried about the location of the fracture.

Diverticulitis – antibiotics, seeds, or exercise

Stollman N, Smalley W, Hirano I, . American Gastroenterological Association Institute Guideline on the Management of Acute Diverticulitis. Gastroenterology. 149(7):1944-9. 2015. PMID: 26453777

This is the new acute diverticulitis guideline from the American Gastroenterological Association Institute (that was as hard to type as it was to read.) I found three of their recommendations interesting:

  • “The AGA suggests that antibiotics should be used selectively, rather than routinely, in patients with acute uncomplicated diverticulitis. (Conditional recommendation, low quality of evidence).” (They note that so far the RCTs showing no benefit of antibiotics have been in inpatients with CT proven diverticulitis.)
  • “The AGA suggests against routinely advising patients with a history of acute diverticulitis to avoid consumption of nuts and popcorn. (Conditional recommendation,very-low quality of evidence).” This is another one of those myths that we breeze over, but can really ruin patients’ quality of life
  • “The AGA suggests advising patients with diverticular disease to consider vigorous physical activity. (Conditional recommendation, very low quality of evidence).” This makes sense, but it has not been part of my discharge script – until now.

People are going to start thinking I have a personal vendetta against antibiotics

Gágyor I, Bleidorn J, Kochen MM, Schmiemann G, Wegscheider K, Hummers-Pradier E. Ibuprofen versus fosfomycin for uncomplicated urinary tract infection in women: randomised controlled trial. BMJ (Clinical research ed.). 351:h6544. 2015. PMID: 26698878 [free full text]

Are antibiotics useful in UTI? I actually think so, but there have been previous studies that illustrate that a lot of UTIs will clear on their own. This was a randomized, double dummy, placebo controlled trial in which 484 women (18-65 years old) received either fosfomycin 3 grams PO or ibuprofen 400mg TID for three days. 69% of the women in the ibuprofen only group had complete resolution of their symptoms, and didn’t use any antibiotics in the next 28 days. That is impressive, but the antibiotics did provide some benefit. The ibuprofen group had more dysuria, based on their definition of ‘non-inferiority’, although the actual numbers for pain look pretty similar. Also there were 5 patients in the ibuprofen group who developed pyelonephritis as compared to only one in the fosfomycin group, although the difference was not statistically significant (p=0.12). I think antibiotics help, but this study reminds us that if you are on the fence, there is no reason to rush the antibiotics. Nearly 7/10 women will clear their UTI without your help. Also, if you call someone back with a positive culture, but they no longer have symptoms, they almost certainly don’t need treatment (assuming they aren’t pregnant).

Bottom line: Antibiotics probably help in UTIs, just not as much as you think

One more time: dex is as good as pred in asthma

Cronin JJ, McCoy S, Kennedy U. A Randomized Trial of Single-Dose Oral Dexamethasone Versus Multidose Prednisolone for Acute Exacerbations of Asthma in Children Who Attend the Emergency Department. Annals of emergency medicine. 2015. PMID: 26460983

I have covered this topic before, but repetition is key in both science and education. This was a randomized, open-label non-inferiority trial comparing a single dose of dexamethasone (0.3mg/kg orally) to prednisolone (1mg/kg PO for 3 days) in 245 children aged 2-16 with known asthma. There was no difference in the primary outcome of PRAM score at day 4 (0.91 versus 0.91; absolute difference 0.005; 95%CI 0.35 to 0.34), although I am not sure this is the most clinically important outcome. There weren’t any differences in the secondary outcomes, such as admission to hospital, length of stay, or return visits.

Bottom line: Once again, dex is great for asthma

Sticking with obvious pediatric topics: ondansetron works

Danewa AS, Shah D, Batra P, Bhattacharya SK, Gupta P. Oral Ondansetron in Management of Dehydrating Diarrhea with Vomiting in Children Aged 3 Months to 5 Years: A Randomized Controlled Trial. The Journal of pediatrics. 169:105-109.e3. 2016. PMID: 26654135

This is another paper I might have skipped because the results seem obvious, but I have recently seen it argued that we use ondansetron too liberally, so I guess it’s worth looking at. This is a well done, double blinded, placebo controlled RCT that enrolled 170 children between 3 months and 5 years of age with acute vomiting and diarrhea and clinical signs of dehydration. Although I worry that the primary outcome of failure of ORT, defined as features of some dehydration after 4 hours of ORT, is a little subjective, the trial was appropriately blinded and placebo controlled. Failure was 31% with ondansetron as compared to 61.5% with placebo, an absolute risk reduction of 30%, or a NNT of about 3. The 30% failure rate does seem high to me though, as I almost never have a kid fail ORT.

Bottom line:  Surprise? Ondansetron does help vomiting kids orally hydrate.

When your heart leaves you speechless

Wasserman JK, Perry JJ, Dowlatshahi D. Isolated transient aphasia at emergency presentation is associated with a high rate of cardioembolic embolism. CJEM. 17(6):624-30. 2015. PMID: 25782453

This is a prospective cohort of 2360 TIA patients, 41 of whom had isolated aphasia at the time of presentation. Patients with isolated aphasia were twice as likely to have a cardiac source of embolism (22.0% vs 10.6%, p=0.037). This is strong, believable data, but I disagree with the authors’ conclusion that “emergency patients with isolated aphasia with a TIA warrant a rapid and thorough assessment for a cardioembolic source”. Non-aphasic patients still had an 11% chance of a cardiac source as compared to 22% with aphasia. Those two numbers clearly necessitate the exact same work up.

Bottom line: This is interesting trivia, but the association of aphasia with cardioembolism is clinically irrelevant.

A Salter Harris Myth Update

Boutis K, Plint A, Stimec J. Radiograph-Negative Lateral Ankle Injuries in Children: Occult Growth Plate Fracture or Sprain? JAMA pediatrics. 170(1):e154114. 2016. PMID: 26747077

Almost everyone has heard my Salter 1 Rant. Here is some more evidence. This is a prospective cohort of 140 children between 5 and 12 years of age with clinically suspected Salter Harris 1 fractures of the ankle. They were all treated with a removable splint (yes – the pediatric tertiary centers are doing this, so you can too). Then all of the children had an MRI at one week. Of the 140 children, 108 had ligamentous injuries on MRI. So take home #1: Despite the old dogma about ligaments being stronger than pediatric bone, children do get ligamentous injuries. Another 27 had isolated bone contusions. Only 4 children (3.0%, 95% CI 0.1-5.9%) actually has Salter Harris 1 fractures, and only 2 of those had any evidence of growth plate injury. And even more important, at 1 month follow up, there was no difference in function between those with MRI confirmed fracture and those without.

Bottom line: Salter Harris 1 fractures are rare and of questionable clinical relevance. Stop casting all these kids.

How important are c-spine precautions in submersion victims?

Watson RS, Cummings P, Quan L, Bratton S, Weiss NS. Cervical spine injuries among submersion victims. The Journal of trauma. 51(4):658-62. 2001. PMID: 11586155

This is a chart review of all submersion victims in the Seattle area between 1974 and 1996. There were a total of 2244 submersion victims, 34% of whom survived until hospital discharge. The prevalence of c-spine injury was 0.49% overall and 0.38% of those who received any medical care (not pronounced dead on scene). All people with c-spine injuries had obvious trauma. (One, for example, was a victim from a plane crash.) The biggest pitfall of this chart review is that someone with a spine injury from submersion might only be coded as a spine injury at discharge, because that was the important injury. These patients would not have been found by the review. However, this isn’t the only reason to be skeptical of cervical collars, so I have no problem removing it if I need better access to a submerged patient’s airway.

Bottom line: A submerged patient is very unlikely to have a c-spine injury if there isn’t obvious signs of trauma

Modified Sgarbossa criteria – now for more than just ECG geeks?

Meyers HP, Limkakeng AT, Jaffa EJ. Validation of the modified Sgarbossa criteria for acute coronary occlusion in the setting of left bundle branch block: A retrospective case-control study. American heart journal. 170(6):1255-64. 2015. PMID: 26678648

This paper is worth a look, if just to review some ECGs. It is a retrospective case-control study looking to validate a modified Sgarbossa rule for diagnosing STEMI in LBBB. This rule uses the ratio of ST elevation to S wave, rather than a set 5mm cut off for the anterior leads. Based on their 258 patients (only 9 with true STEMI), they report a better sensitivity than the original criteria (80% vs 49%, p<0.001) and equal specificity (99% vs 100% p=0.5). I already use these criteria, but I think we should be cautious about the current evidence base. This is retrospective and based on only 9 patients with acute coronary occlusion. More importantly, I wonder about the inter-rater reliability when we are taking multiple measurement in millimetres and dividing them. I already know from reading Dr Smith’s (excellent) blog that he frequently sees small amounts of ST depression that I would have missed or measured differently.

Bottom line: Like many things on the ECG, proportion probably matters, but it isn’t well studied.

Read more on Dr. Smith’s blog here, here, or here.


How many diseases can you diagnose at 20 feet?

Narayana S, McGee S. Bedside Diagnosis of the ‘Red Eye’: A Systematic Review. The American journal of medicine. 128(11):1220-1224.e1. 2015. PMID: 26169885

I’ll just do a very quick note on this systematic review. because I found two numbers interesting. For ruling in “serious eye disease”, photophobia is good (LR+ = 8.3; 95%CI 2.7 – 25.9), but photophobia by indirect illumination (shining the light in the opposite eye) is amazing (LR+ = 28.8; 95%CI 1.8 – 459). The other number I found interesting is that bacterial conjunctivitis can almost be ruled out by “failure to observe a red eye at 20 feet”, although I am not sure there is huge clinical value of differentiating bacterial from viral conjunctivitis.

Bottom line: Worth a read through if you want to better understand your eye exam.

Intralipid review

Hoegberg LC, Bania TC, Lavergne V. Systematic review of the effect of intravenous lipid emulsion therapy for local anesthetic toxicity. Clinical toxicology (Philadelphia, Pa.). 2016. PMID: 26853119

Another quick one: A systematic review of intralipid therapy in local anesthetic toxicity. It might be worth a deep dive, but the quality of the evidence is just so poor that it’s hard to trust any conclusions. For what it is worth, they conclude that intralipid appears effective, but there is no evidence that it is more effective than vasopressors.

My real reason for bringing this up is to lament the quality of toxicology literature in general. I have heard people argue that it would be unethical to randomize these dying patients in order to get good data, but we have to remember that in the absence of good data, the care they are getting is entirely random anyway. The random factor is just the belief of the physician who happens to be on that day. Although these are rare cases, we have the technology to gather data from around the world. We need to do better.

Bottom line: I will probably use intralipid if this comes up, but we really need better science in toxicology.

Osteoarthritis is not an xray diagnosis

Kim C, Nevitt MC, Niu J. Association of hip pain with radiographic evidence of hip osteoarthritis: diagnostic test study. BMJ (Clinical research ed.). 351:h5983. 2015. PMID: 26631296 [free full text]

This study looks at data from 2 large cohort studies: The Framingham study (in which every patient over 50 got a pelvic x-ray, regardless of symptoms) and the osteoarthritis initiative study (which included 4366 patients thought to be at risk for knee arthritis, and again everyone was imaged.) Xray is not predictive of osteoarthritis. In Framingham, only 15.6% of patients with frequent pain (clinical OA) had radiographic evidence of OA and only 20.7% of those patients whose xray indicated OA actually had clinical symptoms. Likewise, In the osteoarthritis initiative study, only 9.1% of patients with symptoms had xray changes, and only 23.8% of patients with xray changes had symptoms.

Bottom line: Xray cannot provide any valuable information about osteoarthritis of the hip

Should we let residents use Google on shift?

Kim S, Noveck H, Galt J, Hogshire L, Willett L, O’Rourke K. Searching for answers to clinical questions using google versus evidence-based summary resources: a randomized controlled crossover study. Academic medicine : journal of the Association of American Medical Colleges. 89(6):940-3. 2014. PMID: 24871247 [free full text]

Rushing around the emergency department, it is obviously tempting to just google something rather than find a specific medical resource, but how good is google? This is a prospective, randomized, controlled, crossover study in which they took 48 internal medicine residents and asked them to answer a series of medical questions. They were randomized to answer 5 questions, either using Google or using their choice of DynaMed, First Consult, or Essential Evidence Plus. They then ‘crossed over’ and answered another 5 questions using the opposite tool. This was repeated for 48 weeks. There was no difference in time to correct answer, response rate, or accuracy. They found answers for 80% of the questions, but the correct answer in only 60%.

Bottom line: Google doesn’t look worse than these specific medical tools, but I really want my residents to be right more than 60% of the time in an open book test.

Cheesy Joke of the Month

What did the pirate say on his 80th birthday?

Aye Matey


#FOAMed of the Month

I often lament the current state of medical science. Data is unreported. Secondary outcomes are reported as primary. Harm outcomes aren’t even mentioned.

COMPare (CEBM Outcome Monitoring Project) is a group of people trying to fix this. You can read a short blog post about it here. In short, they compare publications with the original trial protocol, report discrepancies, write letters to the editors, and report on their progress. It’s an interesting project that is worth checking out.


However, I guess that’s not really education, so I will add a second #FOAMed selection:

Have ever heard of BRASH syndrome? You’ve probably seen it, but if you are like me, you had probably never heard of it before this month:

BRASH syndrome on PulmCrit