The case: A 13 month old presents to the ED with a harsh, barky cough. Parents report a few days of prodromal upper respiratory symptoms but the child has good oral intake and vital signs. The child is playful, with a snotty nose and intermittent stridor. Given the clinical diagnosis of croup, how much dexamethasone do you give? These bits of Free Open Access Medical education (FOAM) served as the impetus behind the development of my steroid in croup practice: EM PEM podcast, HQMedEd video, Emergency Medicine Literature of Note post
Croup (laryngotracheobronchitis)- A single dose of steroids in croup is the mainstay of treatment in mild to moderate disease as they decrease return visits (RR 0.5, 95%CI 0.3-0.7) and hospital length of stay compared with placebo . The number needed to treat (NNT) is 11 in the ED population (or 5 for improvement in a clinical score). The historically touted dose of dexamethasone is 0.6mg/kg orally as this is the most studied dose; however, lower doses have demonstrated non-inferiority and have even made it into textbook recommendations [2,3]. The 0.6mg/kg dose of dexamethasone is equivalent to ~3.75mg/kg of prednisone - overkill for this disease process?
- Dexamethasone 0.15mg/kg PO (can give IV solution orally)
Limitations - the studies are all pretty small and are not powered to detect small differences. Are these small differences clinically significant? Croup also tends to be a self-limiting disease so these studies may be far underpowered to capture rare complications. Furthermore, many of the studies only look at mild to moderate croup so selection bias may demonstrate. Yet Fifoot et al excluded this group (Westley score <2) and still found that 0.15mg/kg was as efficacious as larger doses. The ToPDog study is pending in Australia to demonstrate the effective dose on a larger scale.
Asthma - Glucocorticoids, given within the first hour of presentation to the ED, may reduce need for hospitalization with a NNT of 8 to prevent one hospital admission and, in pediatrics, the length of stay . The literature demonstrates that IV steroids have no superiority over the oral route and the American Thoracic Society recommends oral steroids [8,9]. In fact, Rosen's strongly emphasizes the preference of the oral route, unless the patient is vomiting, in extremis, or has gastrointestinal malabsorption . Prednisone can be swallowed quickly between/during albuterol treatments.
- Pediatrics: Prednisone 1-2 mg/kg PO
- Dexamethasone phosphate 0.6 mg/kg PO
- Adults: Prednisone 40-80 mg/day PO
- Unable to take PO? Methylprednisolone is weight based, at 1mg/kg, not the standard 125 mg given routinely in the ED, which is equivalent to approximately 156 mg of prednisone.
- Prednisone orally, No benefit to doses higher than 60 mg.
- 5 days equivalent to longer courses .
The core emergency medicine texts, Rosen's and Tintinalli's, both recommend IV methylprednisolone in anaphylaxis, followed by a short burst of prednisone for 3-5 days but these recommendations do not come with supporting citations [2,7]. A Cochrane Review turned up no evidence on this therapeutic endeavor . The pathophysiologic argument for IV steroids exists in the notion that patients may have impaired absorption secondary to shunting of blood away from the GI tract. So, what's the answer?
- Shock/sick patients - intravenous steroids such as hydrocortisone or methylprednisolone are likely the right answer.
- The rest of the anaphylaxis patients that are able to swallow - oral prednisone 40-60 mg/day for 3-5 days.
11. Gaeta TJ, Clark S, Pelletier AJ, Camargo CA. National study of US emergency department visits for acute allergic reactions, 1993 to 2004.Ann Allergy Asthma Immunol. 2007 Apr;98(4):360-5.
Clinical Experience of Life-threatening Dabigatran-Related Bleeding at a Large, Tertiary Care, Academic Medical Center: a Case Series. Ross B et al. J Med Toxicol 2014 Jan 3 [Epub ahead of print]
The authors retrospectively reviewed patients on dabigatran (Pradaxa) admitted to UMass Memorial Medical Center (Worcester) for bleeding.
They identified 7 patients with intracranial hemorrhage (ICH) and 4 patients with gastrointestinal bleeding. Although they present copious information regarding baseline characteristics, treatment modalities and outcomes, with a study population this small and non-standardized clinical management, the details represent essentially random noise.
The authors really don’t reach any useful conclusions from the data, and neither can I. Of note (or maybe not), 2 patients were treated with recombinant factor VIIa (rFVIIa) which, as the authors point out, “has not been prospectively studied in humans for the management of dabigatran-related bleeding.” Those 2 patients both had ICH and were the only ones who received rFVIIa. Of course, this association does not establish causation. However, since rFVIIa does not have an established benefit in these patients, but does increase the risk of thrombosis, I wish the authors had specifically discussed these cases. They state that “No patients in this observational case series suffered from any thrombotic event,” they do not explain the basis of this claim.
By the way, there were 3 deaths in this series, 2 in patients with ICH, 1 in a patient with GI bleeding.
Given the weakness of the data, the conclusion is somewhat lame: “Patients with severe dabigatran-related bleeding may benefit from a standardized approach to treatment.” Well, yes. They may benefit from any number of things. Unfortunately, the paper gives little insight into what this “standardized approach” might be.
This post came from a question received in our simulation lab a couple of weeks ago. Mainly, there was a simulated patient with latrodectus envenomation, and there was a fair amount of discussion about skin testing prior to administration on antivenom. Now, when I say discussion, what really happened was some faculty said skin testing was recommended by the package insert and all of their prior readings, whereas the residents were simply asking “why?” Then both groups practiced their google-fu and were able to come up with abstracts to support their viewpoints.
So, yet again, it seems there might be a generational gap between evidence and practice, so I figured I would try to answer their question here. And yet, when I went to search, there haven’t been a large amount of RCTs for skin testing, which isn’t shocking for the toxicology literature.
However, there have been a few decent case series that do not show a benefit to skin testing, as well as a few case series that demonstrate the safety profile of latrodectus antivenom. Putting these together, one could logically make the case against skin testing for latrodectus antivenom. However, there have now been two case reports of deaths from latrodectus antivenom use, one in a young woman with a history of asthma who received an undiluted push dose of antivenom, and more recently a man, also with asthma, who received a diluted dose of antivenom but died after experiencing anaphylaxis. The authors feel his death was likely from PE, but it still happened secondary to antivenom.
This article from Thailand was a retrospective review of snake bites who received antivenom. Over a little more than nine years, there were a total of 254 cases, 211 of which received skin testing. Ten of these patients had positive skin tests, and received different treatment. Desensitization was used in 5, and “close observation” was used in the other 5, but they still received undiluted antivenom. There were no reactions in any of the 10 patients with positive skin tests. Conversely, 7 patients with negative skin tests had reactions to the antivenom, and two who did not receive skin testing also had reactions. So the sensitivity of skin testing in their paper is 0%, and the specificity was 96.4%. Not terribly helpful for making decisions in management.
The good news for the practicing physician is that the weight of the current evidence has led the WHO to recommend against skin testing (at least for snake antivenom) as it leads to delays in treatment and does not help in decision making.
Skin and conjunctival “hypersensitivity” tests will reveal IgE mediated Type I hypersensitivity to horse or sheep proteins. However, since the majority of early (anaphylactic) or late (serum sickness type) antivenom reactions result from direct complement activation rather than from IgE mediated hypersensitivity, these tests are not predictive. Since they may delay treatment and can in themselves be sensitising, these tests should not be used [level of evidence T].
They do have the caveat that they only recommend antivenom treatment in patients who the benefits of said treatment outweigh the risks of allergic reactions.
My personal practice is to not perform skin testing. There is a very small number of patients who I feel need antivenom that I would withhold treatment based on a positive skin test. And since preventive treatment has not been proven effective, again it would only serve to delay definitive care. If I had a patient with a known severe allergy history, I would probably pretreat them (ore concurrently treat them), but I would also get epinephrine and advanced airway equipment to the bedside. The harms and costs of a single dose of steroids and antihistamines are exceedingly low, and you have less of a risk of some “expert” saying you were acting cavalier.
Low incidence of early reactions to horse-derived F(ab′)2 antivenom for snakebites in Thailand
Of note, skin testing doesn’t appear to work for drugs either, so maybe there’s no point in doing it for anything emergent. In this paper, the skin test for cephalosporins had a sensitivity of 0%, specificity of 97.5%, negative predictive value of 99.7%, and a positive predictive value of 0%. Nobody with positive skin tests reacted to the medication, and 4 people with negative tests did have immediate reactions.
Validation of the cephalosporin intradermal skin test for predicting immediate hypersensitivity: a prospective study with drug challenge.