Articles of the month (September 2017)

Back with another edition of the “articles of the every month or so”. I love reading these papers and sharing them with everyone, but I’ve decided I should have at least a modicum of a pathetic social life as well, so every 2 months is probably the best I can do. Even then, I’m not sure they are worth what you are paying for them….

We should be flushing our patients

Driver BE et al. Flush Rate Oxygen for Emergency Airway Preoxygenation. Annals of emergency medicine. 2017; 69(1):1-6. PMID: 27522310

Throughout my training in emergency medicine, including my time on anesthesia, my training in pre-oxygenation was simple: but on a non rebreather and let the patient breath comfortably for 4 minutes or tell them to take 8 full capacity breaths. There are a number of reasons to think this approach is inadequate, especially in critically ill patients who are breathing at high minute ventilations. This is a study looking at 26 healthy volunteers, so there will obviously be some questions about its external validity, but I think the insights are important, because the numbers will be worse in sick emergency department patients. They compared 4 different pre-oxygenation strategies, using the study subject as their own control. They used the standard non-rebreather at 15 L/min; a bag-valve mask at 15 L/min; a non-rebreather at “flush rate oxygen”, meaning the flow was turned up as high as the valve would allow; and a standard face mask at “flush rate”. The outcome was the fraction of expired oxygen at the end of 3 minutes. (We would ideally like the lungs to contain 100% oxygen, allowing for the longest possible time until desaturation.) The key result here is that the standard technique I was taught throughout residency – the non rebreather at 15 L/min – is clearly the worst, with an FeO2 of only 54%. The non-rebreather at flush rate oxygen was the best (FeO2 86%), even beating out the BVM. As a side note, not all BVMs are created equal. Many have an open port near the mask so that if the patient is breathing spontaneously they will entrain room air, and won’t end up breathing anything close to 100% oxygen. (The BVMs where I work are like this. If you want to get 100% oxygen with our BVMs you need to put on a PEEP valve).

Bottom line: When you are pre-oxygenating, turn the oxygen up as high as it will go. (I also place nasal prongs on under the non-rebreather as well.

Push dose problems

Acquisto NM, Bodkin RP, Johnstone C. Medication errors with push dose pressors in the emergency department and intensive care units. The American journal of emergency medicine. 2017. PMID: 28625533

This has been a relatively contentious topic, that has already been discussed by many folks. A discussion on EMCrit can be found here. This is a letter describing a few cases of drug errors in the use of push dose pressers. They decry indication creep (reasonably) and blame FOAM for an increasing number of cases (with no evidence for that claim). My take: these types of errors were common before FOAM. The foam resources I know, especially EMCrit, emphasize using a single technique, over-learning that technique so mistakes aren’t made under pressure, and using drug labels. FOAM resources also offer a rapid mechanism to review push dose pressors in the moment. Compare this approach to how I was taught: there was still an emphasis on push dose pressors during my ICU and anesthesia rotations, but I was informally taught multiple different techniques at the bedside by multiple different physicians. I used different drugs on different days, never drawn up in the same way, and never labelled. As a resident I would be handed an unlabelled syringe and told “just give a cc or 2 at a time if the pressure is low” and then I would be sent down to the CT scanner with the patient. I have no doubt that the FOAM teaching is safer. They make the valid point that push dose pressors are only temporizing (I would add without proven benefit) and that first line therapy should focus on the underlying cause. I agree with that, but then again, I imagine the FOAM community would as well.

Bottom line: Medications errors are common with sick patients, and we need to be careful, but these authors seem to start with a problem with FOAM and try to build their argument from there. (Yes, I have a bias here).

Swabs: predictable, prodigal, pointless?

Torres J, Avalos N, Echols L, Mongelluzzo J, Rodriguez RM. Low yield of blood and wound cultures in patients with skin and soft-tissue infections. The American journal of emergency medicine. 2017; 35(8):1159-1161. PMID: 28592371

This is a prospective study of 734 adult emergency department patients with skin and soft tissue injections looking at the value of cultures. All patients were being admitted to hospital, so it is a higher risk, sicker group of patients than we usually see. (Although they contradict this inclusion criteria by stating in the results that 33% of patients were admitted). The quick summary of this paper is that only a small percentage of blood or wound cultures were positive. Most importantly, the few cultures that came back positive were positive for exactly what the patient was being treated for: Staph and Strep. These would have all been covered by suggested empiric treatment strategies. The cultures would not have changed management in any way. A few details of the numbers: of 86 patients with blood cultures 6 were positive (7%, 4 MSSA, 2 MRSA). Of the 89 patients with purulent wounds, 44 were swabbed, and 13 were positive (30%, 4 MSSA, 4 MRSA, 3 Strep, and 2 polymicrobial. The most important point is that not a single one of these cultures resulted in a change in antibiotics or management, which makes a lot of sense because these results were easily predicted before the culture was sent.

Bottom line: The vast majority of patients with skin and soft tissue infections do not require any cultures. They just don’t change management. Save the cultures for patients who might have an abnormal organism (salt water exposure) or who might require long term antibiotics (chance of underlying osteomyelitis).

What’s the damage of a day’s delay?

Chiang E, Bee C, Harris GJ, Wells TS. Does delayed repair of eyelid lacerations compromise outcome? The American journal of emergency medicine. 2017. PMID: 28473278

This is a small and imperfect study, but with a useful bit of information. They performed a retrospective chart review of eyelid lacerations treated by their opthalmology service and compared those that were repaired in the first 24 hours (66%) to those who were repaired after 24 hours (34%). There were 143 patients identified over a 38 month period. 36% were children. Overall, the rate of complication was not different between the groups (6.3% in those repaired early vs 2% in those repaired late). Clearly, there are many confounders in this retrospective, nonrandomized cohort. However, the low rate of complication is the delayed group is reassuring. It means that if I see a laceration I am uncomfortable repairing, it is probably OK for me to irrigate it, dress it, and have a surgeon take a look the next day.

Bottom line: Not all lacerations need immediate closure.

Which post arrest patients need the cath lab?

Millin MG, Comer AC, Nable JV. Patients without ST elevation after return of spontaneous circulation may benefit from emergent percutaneous intervention: A systematic review and meta-analysis. Resuscitation. 2016; 108:54-60. PMID: 27640933

This is a paper that I presented at the excellent BEEM conference in Niagara Falls (if you are looking for a warm place to go in February, you could join me in Costa Rica for another BEEM course). It is a systematic review and meta-analysis, with some problems, that tried to answer the question: do post-arrest patients without ST elevation on the ECG benefit from going to the cath lab? They include 10 studies (2084 patients), but all are observational. The key numbers: 42% of patients were sent to the cath lab (vs 93% if you have ST elevation), and of those 32% had a “culprit lesion” (vs 73% if you have ST elevation). Although 32% sounds high, there are a lot of problems with the underlying data. The biggest problem is that observational data can only tell us what we are currently doing, not what we should be doing. We don’t know if these “culprit lesions” are actually important. We don’t know if interventions in these patients will translate into clinically important outcomes. To get that information, we would need a prospective randomized control trial. Another point, raised by Rory Spiegel (who wrote this up for BEEM), is that patients should have a good prognosis for neurologically intact survival if they are going to the cath lab, because stents don’t cure brain death.

Bottom line: At this point, we shouldn’t be sending 100% of these patients to the cath lab, but we need to recognize that a significant minority of patients will have an important coronary lesion, and so we need to use our judgement and call cardiology when we are suspicious of an ischemic etiology.

You gotta beat the stress, fool

Lauria MJ, Gallo IA, Rush S, Brooks J, Spiegel R, Weingart SD. Psychological Skills to Improve Emergency Care Providers’ Performance Under Stress. Annals of emergency medicine. 2017. PMID: 28460863

I just wanted to include this to give a shout to Mike Lauria for a great paper on what I think is a really important topics. This is a great, concise alternative if you don’t want to spend a week sorting through my tome on the topic. The quick summary is to remember to “Beat The Stress Fool”, translatied: to breath, talk (self talk), see (visualize), and focus (with a trigger word).

Bottom line: High pressure situations are part of our job. We need to develop skills to manage that pressure.

Tubes for tots?

Steele DW, Adam GP, Di M, Halladay CH, Balk EM, Trikalinos TA. Effectiveness of Tympanostomy Tubes for Otitis Media: A Meta-analysis. Pediatrics. 2017; 139(6). PMID: 28562283

Evidence based medicine telling us that an established practice doesn’t work? Anyone surprised? This is a systematic review and network meta-analysis looking primarily at 16 RCTs. Although tympanostomy tubes seemed to improve hearing between 1 and 3 months after the procedure, there was no long term improvement between 12 and 24 months (as compared to no treatment). More important than the hearing test results, there is no improvement in cognitive, language, and behavioral outcomes. A transient non-clinical benefit doesn’t seem to be worth the risks of surgery to me. Although we don’t make this decision in the emergency department, we do see a lot of otitis media, and this is just one more reason that we should be counselling parents that this is a benign disorder that mostly doesn’t require intervention.

Bottom line: There doesn’t seem to be an real clinical benefit from tympanostomy tubes.

What’s in a number?

Roberts JR, Dollard D. Alcohol levels do not accurately predict physical or mental impairment in ethanol-tolerant subjects: relevance to emergency medicine and dram shop laws. Journal of medical toxicology. 2010; 6(4):438-42. PMID: 20358415

Just a quick reminder that alcohol levels don’t tell us a lot about clinical sobriety. (I’m not sure I need to remind emergency physicians of that fact.) This is just a case report of a 35 year old male brought in with a seizure and ultimate found to have an alcohol level of 515 mg/dL (112 mmol/L). At the time this blood was drawn, the patient walked 30 feet to the bathroom without any gait disturbance. A second blood draw was done to confirm the value and 2 emergency physicians examined the patient and noted no signs of intoxication, altered mental status, or neurologic deficit. Despite being completely clinically sober, the patient was not allowed to leave because of the lab abnormalities. This case report reminds us that chronic users can develop significant tolerance to ethanol and that alcohol levels are not helpful when determining sobriety. Although I would ensure this patient was not driving, I would have allowed him to go home (unless he wanted help with alcohol abuse). In fact, forcing him to stay in hospital probably just increases his risk of alcohol withdrawal.

Bottom line: As a general rule in medicine, treat the patient not the number.  

Haldol for HUGS?

Ramirez R, Stalcup P, Croft B, Darracq MA. Haloperidol undermining gastroparesis symptoms (HUGS) in the emergency department. The American journal of emergency medicine. 2017; 35(8):1118-1120. PMID: 28320545

Cyclic vomiting, gastroparesis, cannabinoid hyperemesis: are there any emergency providers out there who love looking after these conditions? The agent with the best evidence for nausea and vomiting in the emergency department is droperidol – a drug that has never been available to me. However, haloperidol is pharmacologically similar enough that I have been using it fairly routinely for these vomiting syndromes for a few years. This is the first study that I am aware of to provide any evidence for that practice. This is a chart review that identified 52 adult patients (mean age 32) with diabetic gastroparesis who had been treated with haloperidol (5 mg IM). As a control, they looked at the patients’ most recent visit for nausea and vomiting in which they weren’t given haloperidol. When given haloperidol patients were admitted to the hospital less often (10% vs 27%, p = 0.02) and required less pain medication (median 6.75 vs 10.75 morphine equivalents, p=0.009). There was no difference in ED length of stay or hospital length of stay if admitted. There were no episodes of dystonia, akasthesia, excessive sedation, or cardiovascular complications. Although they gave the haloperidol IM here, I tend to use 2.5-5 mg IV.

Bottom line: I would use droperidol if you have it. If you have any power to bring droperidol back to North America, please do. Otherwise, I think haloperidol is a reasonable but unproven options for these difficult to manage patients.

Cheesy Joke of the Month