Welcome to another edition of my favorite emergency medicine articles of the month. Once again, there will be an accompanying podcast with the talented and insightful Dr. Casey Parker on the BroomeDocs website where we briefly discuss these articles.
What carries you across the threshold?
Taylor RA, Singh Gill H, Marcolini EG, Meyers HP, Faust JS, Newman DH.Determination of a Testing Threshold for Lumbar Puncture in the Diagnosis of Subarachnoid Hemorrhage after a Negative Head CT: A Decision Analysis.Academic emergency medicine. 2016. PMID: 27378053
This paper uses computer modelling of the various possible outcomes of a subarachnoid hemorrhage diagnostic algorithm in an attempt to weigh the harms of testing against the benefits of diagnosis and arrive at a reasonable test threshold for lumbar puncture. They come up with a test threshold of 4.3% (range 1.9-9.3%) to warrant an LP after a negative CT. There are a lot of assumptions that go into this conclusion. The least certain assumptions based on the current literature are the morality after a missed SAH and the rate of aneurysmal vs non-aneurysmal bleeds. Both of these, if wrong, would actually tend to increase the test threshold further. However, the biggest harm in these calculations is contrast induced nephropathy and there is some debate about whether that entity even exists. All that being said, 4% is a long way from the 1/700 (or lower) true positive LP rate after a negative CT. (Based on figure 3 in this paper, only patients with a pre-CT pretest probability of 50% or higher who have their CT 156 hours or more after symptom onset would be above the test threshold and therefore warrant LP.) I really love the concept of test thresholds and think we should be talking about them more in medicine. However, there is one major problem with the idea. Benefits and harms can’t easily be calculated down into a single number. Different outcomes have different values. How do you weigh a saved life against a lifetime of dialysis?
Bottom line: This is another nail in the coffin for LP for SAH (but don’t forget, there are still other causes of that headache that might warrant an LP)
Bonus paper: The same topic was also covered in another paper published in the same issue of AEM, with essentially an identical threshold:
Carpenter CR, Hussain AM, Ward MJ. Spontaneous Subarachnoid Hemorrhage: A Systematic Review and Meta-Analysis Describing the Diagnostic Accuracy of History, Physical Exam, Imaging, and Lumbar Puncture with an Exploration of Test Thresholds. Academic emergency medicine. 2016. PMID: 27306497
This second paper is also the best summary I have ever seen on everything you could possibly want to know about the evidence based diagnosis of SAH. It is long, but definitely worth a read.
Is medicine as good as you think it is?
Hoffmann TC, Del Mar C. Patients’ expectations of the benefits and harms of treatments, screening, and tests: a systematic review. JAMA internal medicine. 175(2):274-86. 2015. PMID: 25531451
This is an interesting systematic review. They looked for studies where patients were asked to estimate either the benefit or harm of the intervention (test or treatment) that they were undergoing. They only included studies that looked at quantitative estimates (so no “a lot” or “very little”). They did a very thorough search, but the studies varied significantly, so that a single numerical summary or meta-analysis wasn’t appropriate. In about ⅔ of the studies looking at benefit, more than half of patients significantly over-estimated the benefit they were receiving as compared the true number. Patients similarly under-estimated harm. From this overview it is not possible to determine exactly how large those misunderstandings were or how they might have affected patient decision making. My guess is that even trained physicians often fall into these same biases – overestimating the benefits and underestimating the harms of modern medicine – although probably to a lesser extent. (Otherwise why would anyone still prescribe tamsulosin for kidney stones; or order stress tests; or cast Salter-Harris one injuries?)
Bottom line: We all need to be EBM experts. It’s not good enough to say this medications is good. We need to be providing our patients with accurate NNTs and NNHs
Is the ketofol fad coming to an end?
Ferguson I, Bell A, Treston G, New L, Ding M, Holdgate A. Propofol or Ketofol for Procedural Sedation and Analgesia in Emergency Medicine-The POKER Study: A Randomized Double-Blind Clinical Trial. Annals of emergency medicine. 2016.PMID: 27460905
I will admit my bias from the start: I have never been sold on using two medications when one would do. With ketofol in particular, I seemed to get the downsides of both drugs without any great upside. This is a multicenter, randomized, double-blind trial comparing propofol alone to ketamine plus propofol in 573 adult patients requiring deep procedural sedation in the emergency department. The primary outcome was adverse respiratory outcomes (saturation <94%, apnea for 15 seconds, hypoventilation <9 breaths/min, laryngospasm, and aspiration). Although those sound reasonable, I question their cutoffs – I don’t care if my patient has sats of 91% or is only breathing 6 times a minute, assuming it is short lived, as it always is with propofol. You might also question whether this was a fair comparator, as apnea and hypoventilation are known side effects of propofol, but are rarely important, and one of the reasons we choose ketamine as an alternative. That being said, it made no difference, as this was a negative trial, with 7% of the ketofol group and 9% of the propofol group having the primary outcomes (absolute difference 2%, 95%CI -2 to 6%). As for secondary outcomes, not surprisingly there was more hypotension in the propofol group (8% vs 1%) and more agitation in the ketofol group (5% vs 2%), but I don’t think any of these events were clinically important from their description. All patients were completely satisfied with their sedation. This is the best study of ketofol that we have so far, but the are a number of limitations. This was a convenience sample and a large number of patients were excluded because of “physician preference”, so selection bias is quite likely. Ketofol fans might note that they only used one mixture of ketamine and propofol (1:1) and that there may be better ways to dose the two drugs. The biggest problem is that true adverse events are so rare in procedural sedation that unless you run a massive trial, every comparison is going to look negative. (For example, there were 0 cases of airway obstruction, laryngospasm, or aspiration in this study, so we really have no idea which medication is better.) There may be times to use both of these medications in a sedation, but for the vast majority of patients choosing the single most appropriate drug makes sense.
Bottom line: There is no benefit in routinely combining ketamine and propofol for sedation
Staged repetition: Anti-NMDA encephalitis is an important diagnosis we must recognize
Lasoff D, Vilke G, Nordstrom K, Wilson M. Psychiatric Emergencies for Clinicians: Detection and Management of Anti-N-Methyl-D-Asparate Receptor Encephalitis. The Journal of emergency medicine. 2016. PMID: 27431869
I have included this topic before, but as a new diagnosis, it will probably take a while before it is fully ingrained into our consciousness. NMDA receptor antagonist encephalitis is probably more common than HSV encephalitis, but it is definitely diagnosed less. As many as 1% of all ICU admissions might have these antibodies, explaining a large number of unexplained encephalitis cases. Unfortunately, there is no aspect of the history or physical that is specific. These patients generally progress through 4 phases: 1) general viral prodrome, 2) psychiatric symptoms (agitation, hallucinations), 3) neurologic symptoms (seizures, confusion, movement disorders), and 4) autonomic instability. No imaging test is all that helpful. The test of choice is CSF looking for the anti-NMDA antibodies. In the ED, management is going to be symptomatic, but it is essential that we think about the condition, so that IVIG can be started if the CSF tests positive.
Bottom line: Anti-NMDA encephalitis is an important diagnosis to keep in mind.
The one time that lytics aren’t controversial
Sharifi M, Berger J, Beeston P. Pulseless electrical activity in pulmonary embolism treated with thrombolysis (from the “PEAPETT” study). The American journal of emergency medicine. 2016. PMID: 27422214
This is a retrospective study looking at a group of 23 patients who had all been referred to a specialist cardiovascular team after a PEA cardiovascular arrest secondary to PE (20 diagnosed by CT and 3 by right heart strain on bedside echo) and who received tPA. 4 of the 20 CTs were done after the arrest and tPA had been given. The patients all received 50mg of tPA followed by heparin. Return of spontaneous circulation and hemodynamic stability was achieved within 15 minutes for 22 of the 23 patients. At 22 month follow-up, 20 of 23 patients were still alive. There were no major or minor bleeding events recorded. This group was relatively sick, with a mean age of 72, lots of comorbidities, and 22% having a cancer history. Of course, there is no comparison group and this is a select group of patients that was actually referred, so I wouldn’t pay too much attention to the specific numbers, but these outcomes are way better than normal PEA.
Bottom line: This observational data indicates excellent outcomes if you are able to identify and rapidly treat PE as the cause of PEA. This is probably a good reason to become very comfortable with bedside echo – it might really make a difference.
How do you know your trauma patient is a nurse? They always have an empty stomach and a full bladder
Recio-Saucedo A, Pope C, Dall’Ora C. Safe staffing for nursing in emergency departments: evidence review. Emergency medicine journal : EMJ. 32(11):888-94. 2015. PMID: 26273096
This is a review of studies that looked at the impacts of different nursing ratios on emergency department care. Higher patient to nurse ratios were associated with higher numbers of patients leaving without being seen, longer patient times in the ED, and lower patient satisfaction. Nursing ratios higher than 1:1 for resuscitation, 2:1 for critical patients, and 4:1 for all other ED patients increased total time in the ED and delayed admission to the inpatient wards. There was no association between nursing level and medication errors or critical interventions like time to antibiotics, which is indicative of the high level of dedication and skill of our nurses, but I can’t imagine that errors don’t increase with higher ratios. This paper doesn’t have all the answers, but I include it because it is a very important topic that needs attention. My sense is that we are constantly pushing our nurses to cover more patients and I wonder about the breaking point. It is not uncommon to see nurses miss their breaks because there are just too many patients to cover.
Bottom line: Good luck getting your hospital to fund enough nurses – but maybe the higher rate of left without being seen and lower satisfaction scores might help convince your administrators
Do you really expect drugs to work after you’re dead?
Evans ME, Chassee T. BET 2: Usefulness of epinephrine in out-of-hospital cardiac arrest. Emergency medicine journal : EMJ. 33(5):367-8. 2016. PMID:27099380
One of the Best BET series: epinephrine in out of hospital cardiac arrest. There is nothing that we didn’t know here – epinephrine seems to increase ROSC and survival to admission, but without changing survival to discharge or good neurologic outcomes. That has always sounded like a harm to me, rather than a benefit. They didn’t identify the Jacobs trial though, which makes me somewhat concerned about the quality of the search. Is this because we use 2 names for the same drug (epinephrine and adrenaline)? I won’t say which name is wrong, as my new podcast partner is an Australian, but it does point to a bigger issue in medicine: there is no reason that any medicine should be referred to by more than one name. We really need to ban brand names and simplify the current mess of terminology.
Bottom line: Just a reminder of the lack of evidence for epinephrine while we all eagerly await the results of PARAMEDIC 2
Are we going to find any drugs that bring people back to life?
Lee YH, Lee KJ, Min YH. Refractory ventricular fibrillation treated with esmolol.Resuscitation. 2016. PMID: 27523955
I had previously included a retrospective chart review that showed an association between esmolol use in refractory ventricular fibrillation and survival to hospital discharge. (January 2015) However, here is a different retrospective study of esmolol that looks less promising. This is a retrospective look at a period before and after the introduction of esmolol into the algorithm for refractory v.fib for this service. Out of a total of 383 out of hospital cardiac arrests, 183 presented with v.fib, and 41 had refractory v.fib. There was more ROSC in the esmolol group (56 vs 16% p=0.007) but there was not a significantly different rate of long term survival or survival with good neurologic outcome. To me, any drug that puts more patients in the ICU with no hope of leaving is not just ineffective, but clearly harmful. Of course, this is a small retrospective study, but the before and after design probably produces fewer biases than the previous chart review.
Bottom line: This may be reasonable in a very select population where ROSC might get them to definitive therapy, but I won’t be using it routinely yet
Face mask, N95, full SCUBA gear?
Smith JD, MacDougall CC, Johnstone J, Copes RA, Schwartz B, Garber GE.Effectiveness of N95 respirators versus surgical masks in protecting healthcare workers from acute respiratory infection: a systematic review and meta-analysis. CMAJ : Canadian Medical Association journal. 188(8):567-74. 2016. PMID: 26952529 [free full text]
This systematic review identified 29 studies, 6 of which were appropriate for the quantitative meta-analysis. N95 masks did not appear to be protective in any of the clinical outcomes: influenza like illness, lab confirmed respiratory infection, or days off work. However, the masks do seem to work in laboratory settings, so perhaps the reason that they don’t work in the real world is that we don’t wear them properly? Also, the confidence intervals here are relatively large and all the point estimates come down on the side of N95s being better. Given the extra expense and annoyance, this probably warrants a high quality large RCT, but for now the best answer is simply: we don’t know.
Bottom line: N95 masks might not be any better than surgical masks, but we really can’t be sure at this point
Least clinically relevant paper of the month (unless you work next to an amusement park)
Rietveld S, van Beest I. Rollercoaster asthma: when positive emotional stress interferes with dyspnea perception. Behaviour research and therapy. 45(5):977-87. 2007. PMID: 16989773
I love this paper, but I am not sure why. I think it has just been too long since I was last on a rollercoaster. This study explored the relationship between perceived dyspnea (subjective) and objective lung function based on emotional stress. Prior research has shown that negative stress or emotions are associated with an increased sensation of dyspnea despite unchanged objective respiratory values. The authors wondered whether the opposite – positive stress, or joyful situations – would decrease the sensation of dyspnea. They provoked this sensation by having women with severe asthma ride rollercoasters. (They prescreened the women to ensure that they all liked rollercoasters.) Immediately after the ride, at a time of positive stress, perceived dyspnea was significantly lower, despite FEV1 measurements that were actually a little lower than baseline. In other words, positive emotions seem to decrease the sensation of dyspnea.
Bottom line: Dyspnea is a complex phenomenon. Although I don’t have a rollercoaster in my department, an awareness of the impacts of emotion on dyspnea could help us improve patient comfort.
If the rollercoaster didn’t work, maybe it’s time for apneic oxygenation?
Denton G, Howard L. BET 1: Does apnoeic oxygenation reduce the risk of desaturation in patients requiring endotracheal intubation? Emergency medicine journal : EMJ. 33(7):517-9. 2016. PMID: 27330180
I have been a fan of apneic oxygenation since I first heard of it, despite very limited evidence for its use in the emergency department. This is a review for the best BETs series and they conclude that the evidence favours using apneic oxygenation in the emergency department. The majority of the papers they reviewed demonstrated decreased desaturation with apneic oxygenation and the intervention is essentially harmless (assuming you have an extra oxygen port). Of course, they conclude that more and better quality research is necessary. I have been using this strategy despite the lack of emergency department evidence because there is actually decent evidence in the anesthesia literature and the physiologic evidence is strong. I would love to see one really good RCT to settle the issue, but I think the literature to date definitely supports using apneic oxygenation routinely, unless it would require a huge system change to implement (such as for EMS folks who only have one oxygen source).
Bottom line: I will continue to use apneic oxygenation routinely during all my intubations
Red bloods cells can also be fashionably late – when anticoagulated
Chauny JM, Marquis M, Bernard F. Risk of Delayed Intracranial Hemorrhage in Anticoagulated Patients with Mild Traumatic Brain Injury: Systematic Review and Meta-Analysis. The Journal of emergency medicine. 2016. PMID: 27473443
We all know about the potential for a delayed intracranial bleed in patients on anticoagulants, but no one seems sure what to do with that information. Personally, I have always believed that symptoms will identify clinically important bleeds (which is probably true for these patients on arrival as well.) This is a systematic review and meta-analysis of 7 studies including 1594 anticoagulated patients who were rescanned after an initially negative CT. The overall rate of positive CT at 24 hours was 0.6% and the chance of neurosurgical intervention or death was 0.1%.
Bottom line: The risk of delayed hemorrhage in anticoagulated patients, though real, is not large enough to warrant admission or routine repeat CTs. Good discharge instructions are all that is necessary.
ABG = A bad gas
Kelly AM. Can VBG analysis replace ABG analysis in emergency care?Emergency medicine journal : EMJ. 33(2):152-4. 2016. PMID: 25552544
most situations we should be using VBGs and not ABGs in the emergency department in these emails before. This is a nice review of the evidence on the topic. Perhaps the most important thing to remember is the downsides of ABGs: pain, increased technical difficulty, and rare but serious adverse events. The evidence indicates that a VBG as good as an ABG for measuring pH and bicarbonate. For pCO2 the agreement is not great, but Dr Kelly makes the important point that the exact numbers are not as important as the ability to answer a clinical question. If the VBG reveals hypercapnia, the number might not match the ABG, but the patient is clearly going to be hypercapneic. Also, a VBG with a pCO2 less than 45 mmHg rules out clinically significant hypercapnia with 100% sensitivity (95% CI 97-100%). You shouldn’t need a blood gas to determine the pO2because you have a sat probe.
No comments:
Post a Comment