top of page


By Charles Rich

On this page we publish the outcome of trainee-led debate about recent publications, developents and ideas being discussed around the EM world.

25th September 2018

The Thames Valley EM trainees recently looked at the following blog: on the BICAR-ICU trial on the use of 4.2% sodium bicarbonate therapy to treat metabolic acidosis in ICU patients by Josh Farkas, assistant professor of pulmonary and critical care medicine, University of Vermont.



The summary of the BICAR-ICU trial


Trial: open-label, multi-centre RCT in France between 2015-2017


Inclusion: <48h in ICU, severe metabolic acidosis with a pH<7.2, qSOFA>3 OR lactate>2

Exclusion: GI/renal bicarbonate loss (renal tubular acidosis, CKD, ketoacidosis, exogenous acid load eg salicylate overdose)


Intervention: 4.2% sodium bicarbonate 125-250ml in 30min, repeated until pH>7.3 up to a max of 1L bicarbonate in 24h (equivalent to 3.2L if using 1.2% bicarb), for 28 days or until discharge

Control: no bicarb 


Baseline characteristics: well-matched, fairly representative (43% surgical, 53% sepsis, 83% intubated, 80% pressors) so good external validity


Primary outcome: composite of all-cause mortality at 28 days or any organ failure at day 7 - this is a confusing composite that nobody really understood the rationale for


Sample size for power calculation based on retrospective work correlating acidosis with poor outcomes (probable over-estimate of effect so under-estimate of sample size needed so underpowered study)


24% controls got bicarbonate so lots of protocol violations.

109 patients were excluded as they had already had bicarbonate so the sample therefore had less people who were likely to benefit


Result was a neutral primary endpoint with a non-significant trend towards good outcomes

Pre-specified subgroup with worse AKI: mortality improved 63% to 46% (p=0.017)


Major secondary endpoint was need for Dialysis: 53% in control group down to 35% in intervention group (p=0.0009) so NNT=6! The avoidance of dialysis was due to less hyperkalaemia, less acidaemia and better urine output.

For those who needed RRT, bicarbonate delayed  this by 12 hours.


Minor secondary endpoint was hyperalaemia 49% down to 32% (p=0.0006).


Study weaknesses: no differentiation between NAGMA and other acidoses (analysis of the subgroup of patients with worse renal function suggest that bicarbonate is good for uraemic acidosis)

Study only assessed 4.2% (hypertonic)

Reported adverse events: increased metabolic alkalosis (authors wonder if this is due to situations where lactic acidosis is treated with bicarb causing a rebound effect; increased hypernatraemia and increased hypocalcaemia but no life-threatening events.


Conclusions: bicarbonate is good for NAGMA and uraemic acidosis but bad for lactic acidosis 


Group Discussion


The post essentially describes metabolic acidosis as 3 distinct entities that vary in their response to sodium bicarbonate treatment.

·        NAGMA, which is essentially bicarbonate deficiency

·        Uraemia where bicarbonate is probably useful in the context of hyperkalaemia

·        Lactic/ketoacidosis: no role for bicarbonate


The blog describes the different concentrations of sodium bicarbonate as very different medicines.

  • 8.4%: no effect on K, increase Na

  • 4.2%: possibly mild decrease in K (from the trial)

  • 1.4%: lower K, increases the volume


The group overall reported a reluctance to give 8.4% peripherally unless peri-arrest.

There was a debate about the purported increased intracellular acidosis that may result from bicarbonate treatment. Trainees with more ICU experience say that 1.26% is used in ICU for renal patients but very carefully. In ED we often defer to intensivist’s experience - these patients are almost universally heading to ICU anyway. What if a mixed picture of acidosis (NAGMA and lactic)? The blog suggests treatment with sodium bicarbonate may have a role by treating the NAGMA component.


Overall an interesting study, an interesting commentary, and an interesting discussion. Above all highlights the feelings amongst EM trainees that sodium bicarbonate therapy is complex and requires an understanding of multiple physiological processes to get right.


Consultant Comment 
by Dr Liza Keating (EM and ICM at Royal Berkshire Hospital)

"Good summary. Personally, I have never been sure whether severe metabolic acidaemia contributes to organ dysfunction OR is a marker of just how sick someone is……and does this matter either way?  Firstly – never used 4.2% so slightly unconventional choice. Also as you say a lot of contamination – ie 24% control got bicarb anyway (!)  Summary: in patients with severe metabolic acidaemia,  bicarb treatment had NO effect on the primary composite outcome (ie, mortality by day 28 or the presence of at least one organ failure at day 7) BUT secondary analysis in the subgroup of patients with AKI bicarb treatment did decrease the composite outcome and 28 day mortality  - don’t forget there is a HIGH risk of  a type I error based on the outlined limitations of the study. SO: this study will reassure those who already use bicarb for correcting metabolic acidaemia, that this may lessen the need for RRT. On the other hand, for those that opt to avoid bicarb, I’m still not clear that this will change anything. I guess one word of caution…..have seen audit data from many series show an incorrect use of NIV for metabolic acidosis so maybe some concerns about around using bicarb for metabolic acidosis – in the heat of the moment before you have had a chance to get all the data you need……surely not you say. One more thing -  they are a little vague about when to start the filter in this non-blinded study…..AND don’t forget the value of the filter in fluid removal. So not really a game changer for me.


9th October 2018

Check out the following blog from earlier this year

by Prof Steve Smith, author of 'Dr Smith's ECG blog' and professor of EM at University of Minnesota

Summary of Blog

The current STEMI criteria only identify 70% of patients with acute coronary occlusions. The first changes of acute coronary occlusion on ECG are hyperacute T waves and if detected provide an opportunity for immediate coronary intervention before significant necrosis. Dr Smith and other prominent EM physicians advocate that the STEMI/NSTEMI duality is out-dated, inappropriate and not evidence-based.

The blog proposes that a new heading is used to diagnose those patients with chest pain in the ED who need to be referred for immediate coronary intervention - 'Occlusive Myocardial Infarction'. Read the blog for a blow-by-blow account of the last 30 years in MI diagnosis and management and why the team think the time is right to update our thinking.

Summary of Group Discussion

The group's main area of debate was on how this particular change on our part would play out in real-life with our cardiology colleagues who may be just as dogmatic about traditional practice. The following are paraphrased from individuals discussing the blog:

"What would a cardiologist say if we asked them to open the cath lab for someone with chest pain and hyperacute T waves but no ST elevation?"

"In the past we've called cardiology about patients who clinically are definitely having an MI but because of no ST elevation they've insisted on waiting for troponins or ECG progression"

"This highlights the ever-increasing issue of advances in emergency medicine only creating further conflict between EM doctors and hospital specialists - what's the point in us knowing this if we can't affect change because definitive care comes from cardiologists, not us"

"Advocating for our patients is even harder where tertiary referral to another centre for coronary intervention is necessary"

"In light of the recent publication of the HI-STEACS trial showing that use of high-sensitivity troponins has made more diagnoses of MI but has not improved outcomes, what is the evidence that PPCI based on 'OMI' criteria as opposed to traditional 'STEMI' criteria improves outcomes for patients"

Consultant Comment
by Dr Francoise Ticehurst, Wexham Park Hospital
This is a very interesting blog and a topic which all emergency physicians will no doubt have debated about - is this really STE or not, does the cardiologist think this is STE, you think they should go for PPCI but the cardiologist doesn't and so on. I think this blog clearly highlights how difficult the diagnosis can be at times. I agree that we really have got hung up on the terms STEMI and NSTEMI and they do truely feel out dated and a move away from this terminology would be a good thing. The crux of the whole debate in the UK, relies on buy in into the OMI concept from our cardiology colleagues. Also we seem to struggle with varying thresholds being given by different PCI centres. The group discussion has highlighted some very real issues and the difficulties in changing clinical practice on a day to day basis. You definitely need your local cardiologists on board with this and maybe a joint meeting, presenting information such as this blog would help bridge the gap around expectations and encourage more collaborative working. I think also teaching to more junior ED doctors needs to emphasise that it's not all about STE and that ECGs need a more careful analysis looking for the hyperacute T waves, which can be challenging in the middle of a busy ED shift. Finally I would say that we as emergency physicians are masters of change and are highly adaptable and must continue to be the advocate for our patients, which may lead to enhancing our powers of persuasion too.

23rd October 2018

The Thames Valley EM trainees discussed the following blog talking about the use of cricoid pressure during RSI

by Simon Carley, EM Professor in Manchester

Summary of Blog

This blog relates to a linked study recently published in JAMA Surgery. This was a large, randomised, multi-centre RCT powered to show that using a sham procedure was non-inferior to using cricoid pressure at reducing rates of aspiration during RSI.

Cricoid pressure (the Sellick manoeuvre) is part of the traditional teaching of RSI but the evidence for it's ability to reduce aspiration is thin. The study randomised nearly 3500 adult patients, deemed high risk of aspiration, undergoing pre-operative RSI, which the blog argues makes this study population fairly representative of ED patients.

The primary outcome, rates of aspiration, was not statistically different between the two groups but as the overall rates were very low (0.5-0.6%) and the power calculation used an expected rate of 2.8% the study is effectively underpowered. The blog makes it clear that the important thing here is not the power but the overall low rates of aspiration. If the study was adequately powered, any difference caused by the use of cricoid would have a very high NNT and thus the study gives us very good information about the non-inferiority of the sham procedure.

The secondary outcome of view at laryngoscopy and time to intubation showed a non-significant trend towards favouring the sham procedure. Overall the blog argues that for all but the highest-risk patients (bowel obstruction, pregnancy) one could make a good argument against the use of cricoid.

Summary of Group Discussion

The group had a variety of points to make about the impact of this study. Along the same lines as Prof Carley the group had concerns about the use of a non-inferiority trial design and the fact that the study was underpowered.

One issue is that when applying cricoid pressure one does not know how much force is being applied and this might explain why rates of aspiration are similar but rates of airway occlusion and view obstruction are higher.

The opinion of doctors working in the pre-hospital environment is that it is considered far less important than other aspects of RSI so that it's use is not considered a priority, particularly where staff numbers are short.

Many doctors said that where they work all drug-assisted intubations are done by the anaesthetic/ICU teams and so they are de-skilled and typically just defer to the specialist teams. Others were concerned that gudielines that advocate cricoid for all RSIs hamper people from deciding not to use it.

Overall an engaging discussion that reminds us of just how complicated common procedures are, how much variation there is in practice and how variable the evidence is.

6th November 2018

Dynamic resuscitation is the theme of this blog by Dr Peter Kas

Summary of Blog

This article reports on  the growing body of evidence that the management of cardiac arrest can be tailored to an individual patient, in this case by dosing adrenaline to achieve a desired diastolic blood pressure, to ensure ability to perfuse coronary arteries. Current practice around the region follows standard ALS guidelines. The evidence for its use is weak, indeed the recently-published PARAMEDIC-2 trial suggested that whilst rates of survival were very slightly better, rates of survival with favourable neurological outcome were worse in patients given adrenaline.


Increasing afterload by giving adrenaline can lead to loss of circulation. A diastolic BP of 25-35mmHg is known to be adequate for cerebral and coronary perfusion, which should be the aim of giving the drug. This can only be done accurately using an arterial line, which can be inserted under ultrasound-guidance at the first rhythm check but requires practice and co-ordination. The blog offers the following guidance and although other suggestions are used elsewhere, the principles are worth examining.


  • If the heart is seen to be beating on echo and the diastolic BP is >40mmHg don't give adrenaline.

  • If the heart is beating but the diastolic BP is low give 30-50mcg of adrenaline then re-assess.

  • If there is refractory VT/VF then adrenaline will only add to the intrinsic catecholamine surge and so should be stopped.

  • Consider not using adrenaline if the arrest has been over 20 minutes in duration or after the first 3 doses as adrenaline has been found to be detrimental in the 'metabolic phase' of resuscitation. 

Summary of Group Discussion

The group found this discussion very engaging. There was a feeling that whilst ALS exists to allow standardised practice of cardiac arrest management for providers of all types in all settings, we have a duty to as expert resuscitationists to push the boundaries and deliver the highest standard of care possible. This of course does not mean that for situations outside the resus room, ALS should not be the pathway to follow. One doctor commented that other environments might feel under pressure to provide a similar standard of care, despite having neither the expertise nor the equipment and resources.


We already perform expert interventions in the resus room for cardiac arrest patients, which exist more as adjuncts to, rather than replacements for the ALS algorithm, such as mechanical CPR and cardiac ultrasound. The group felt that the idea of adapting resuscitation to individual patients made a lot of sense, particularly if that intervention can be done during what the blog describes as the 'circulatory phase' of the arrest rather than the later 'metabolic phase' where adrenaline is thought less likely to be of benefit.

Despite being routine practice in many countries and in a few centres in the UK, the placement of arterial catheters during cardiac arrest, which is necessary for this approach of dynamic resuscitation to guide the titration of adrenaline, will be a culture change to all resus rooms in the region. It will need experienced staff with the opportunity to practice in order to achieve rapid success at insertion, security and transduction. Such staffing is likely to be limited to daytime hours.

Those with subspecialist interest in paediatric EM commented that these approaches would be standard of care in PICU and would provide several advantages if performed in the resus for patients with out-of-hospital cardiac arrest. This includes an assessment of whether or not vasoactive medication is likely to be of benefit and particularly when considering children with known LV failure, whether other agents might be more appropriate.

Consultant Comment

Dr David Carke, Royal Berkshire Hospital

This is interesting stuff but it's important that we express caution before applying this into practice.


Despite the author’s criticism of ALS guidance, there is no doubt that the regulation of arrest protocols had led to a much more professional and co-ordinated team response to incredibly stressful scenarios.  Having experienced the chaos of un-coordinated cardiac arrests on dark elderly care wards as a PRHO – I think it’s important not to jeopardize that team approach; unless of course – there is a very good reason.


It’s difficult to comment on the blog without reading the original articles – but my perception was that there was a lack of quality evidence prior to PARAMEDIC2. This is notable by its absence – and does make me wonder whether literature has been found to support an opinion rather – rather reading literature to form an opinion.


Not giving adrenaline in a refractory VT/VF seems reasonable and more cautious dosing in certain situations in peri-arrest is also intriguing – but that is as far as I’m willing to take it.


As senior ED clinicians our role is to have a greater understanding of ‘cardiac arrests’ as a variety of pathology and not a single entity – and we are beginning to see this in hypovolaeemic cardiac arrests associated with trauma.

Similarly – we need to have a good understanding of the physiology of the cardio-vascular system and then the pharmacological actions we are inflicting upon it.


Good ALS is still about high quality CPR, early defibrillation and reversing the reversible. To suggest that early arterial lines and non-standardized doses of medication should be the new priority doesn’t seem applicable to my practice.


The challenge to us it to have the greater understanding of these different features and strong enough leadership to guide a team on a well-rehearsed track – with the confidence to occasional step off that track – but only if we’re confident it’s the right thing

19th February 2019

The trainees discussed a recent article in EMJ about management of patients with Left Ventricular Assist Devices (LVADs). 

Bowles CT, Hards R, Wrightson N, et al. Algorithms to guide ambulance clinicians in the management of emergencies in patients with implanted rotary left ventricular assist devices. Emerg Med J 2017;34:842-850.

LVADs are being used ever more frequently as a bridge to transplantation in patients with severe heart failure. Many of these devices produce a continuous outflow through the left ventricle and as such these patients will never have a palpable pulse.

Patients will be based close to their local cardiac centre and will have patient-specific troubleshooting plans; however, it is increasingly likely that a patient with an LVAD could end up being brought to a DGH with a critical presentation e.g. unresponsive. Relatives, bystanders, pre-hospital teams and emergency department staff are likely to be unfamiliar with these presentations and will be confused by a well patient with no pulse or a conscious patient in VF.

The article talks through a series of 3 proposed algorithms for approaching the emergency management of these patients

1. Initial assessment of the LVAD patient

2. LVAD troubleshooting

3. Ensuring adequate circulation to sustain life

The trainees recognised the need for basic understanding of this rare but critical scenario and values the availability of a suggested management algorithm. These algorithms are suggested by an experienced group of professionals who are involved in the routine care of LVAD patients but it must be stressed that these algorithms have not been endorsed by RCEM or the Resus Council UK.

2nd April 2019 - a blog by Justin Morgenstern, focussing on the following study and its context - heparin does not have benefit in ACS. Chen JY, He PC, Liu YH et al. Association of Parental Anticoagulation Therapy With Outcomes in Chinese Patients Undergoing Percutaneous Coronary Intervention for Non-ST-Segment Elevation Acute Coronary Syndromes. JAMA internal medicine. 2018; PMID: 30592483

The blog references 7 studies from between 1988 and 1995 (an era before coronary stenting), none of which suggest a benefit from heparin in ACS (beyond a slight reduction in re-infarction that only persists for the time the patient remains in hospital). It then made reference to the study quoted above, which showed no difference in in-hospital mortality in NSTEMI patients treated with heparin compared to those who were not; however, rates of major bleeding were increased.

The Thames Valley EM registrars found it difficult to reconcile these clear results with the 2014 AHA/ACC guidelines that continue to recommend heparin for NSTEMI with a Class 1 recommendation although they do acknowledge that the studies on which they base this are from a different era. In 2011 the ESC guidelines explain that unless the patient is revascularized before heparin treatment is stopped, there is often a recurrence of events. Thus the ESC acknowledge that heparin is a temporizing and not a life-saving measure.

The focus of the debate then turned to a reflection on how difficult it can be to practice evidence-based medicine in a system that relies on protocols, pathways and guidelines. Remarks were made about how it would be dangerous to go against guidelines but it will now be difficult being asked to prescribe heparin to ACS patients when we know there is no benefit but it does increase the risk of harm. 

Should we try to change the guidelines? Should we as an EM co-operative unilaterally take the decision to stop this dubious practice? Would hospital management support this if cardiology refused to agree? Does this decision fall under the auspices of emergency medicine or cardiology and to what extent is our specialty existing to provide assistance to specialists? 

The evidence seems clear but the questions of implementation are difficult. Will the next generation of EM doctors, with the world of FOAMEd at their feet, be able to resolve such problems?

16th April 2019 is a blog by Dr Justin Morgenstern about paediatric UTIs, their significant, their diagnosis and the role of antibiotics in their treatment.

This blog and its multiple linked papers explore this complicated issue and the Thames Valley trainees were challenged to consider in which circumstances they prescribe antibiotics to treat UTIs in children and the reasoning behind these choices. 

This excellent and exhaustive piece of critical appraisal can be summarised by the following main points (either direct quote or paraphrase).

  • If a child is unwell take cultures and treat as appropriate

  • In well-looking children UTI confers a low risk for sepsis and does not seem to cause any long term renal damage

  • In well-looking children there is no difference in outcomes between starting antibiotics at index presentation compared to day 5 of fever

  • There will be a high rate of false positive urine cultures given the high prevalence of asymptomatic bacteriuria

  • Logic suggests waiting until the 5th day of fever before testing for UTI - this will reduce false positive cultures but will not increase harm

  • There is no evidence to support routine imaging of children after a first UTI

One trainee added more useful information by quoting a paediatrician she works with who says that there is no value in performing Dipstix testing on urine from children who are not yet potty trained as part of a work up for possible infection.

We discussed how this data fits with the NICE guidelines on the management of children with fever. Trainees seemed comfortable with the idea of taking a more restrictive approach to the prescription of antibiotics for suspected UTIs in well-looking children without urinary symptoms although there is still uncertainty surrounding the ideal management for infants with fever and bacteriuria.

9th July 2019 - a blog by Jeff Riddell, assistant professor of clinical emergency medicine at UCLA and reviewed by Reuben Strayer, associate medical director, department of emergency medicine, Maimonides medical centre, NY, regarding the tricky subject of rapid tranquilisation in the ED. 

The blog talks about the extremely troublesome group of patients who are severely cerebrally agitated, cannot be talked down, are requiring physical restraint by multiple security or law enforcement personnel and are refusing to comply with even the most basic care interventions. 

The registrars recognise the challenge posed by these patients and the danger they pose to themselves and others. There has also been some confusion regarding best practice in terms of rapid chemical sedation with a disparity between RCEM guidance, local policies and individual consultants' preferences.

The blog points out that these patients often need immediate interventions, e.g. active bleeding from a wound or assessment of reported chest pain or confusion; however, interventions like wound care, clinical examination, recording vital signs, electrocardiography and blood gas analysis are impossible without sedation. Often the patient is wearing a spit-guard and is at risk from positional asphyxia during physical restraint as well as injury to extremities, hyperthermia, hypoxia, acidosis, hyperkalaemia, rhabdomyolysis, seizures, coma and death. There is a need to gain control of the situation as soon as possible for the benefit of the patient, the staff, those restraining the patient and the other patients in the department.


Controversy exists around the use of ketamine for rapid tranquilisation for 2 main reasons. First, there are data from one study, in the pre-hospital setting, where there were a high number of intubations in a group of patients sedated using ketamine. Second, there are concerns about rendering a cerebrally agitated patient into a dissociated state, particularly by medication acting on NMDA receptors when often we are not sure whether stimulation or inhibition of these receptors is responsible for the agitated state in the first place.


RCEM has a policy on rapid tranquilisation ( which lists the time of onset for various agents when given via the IV or (what is mostly relevant in these scenarios) the IM route.


IM Midazolam 5mg: 10-15 minutes

IM lorazepam 4mg: 15-30 minutes

IM diazepam 10mg: 15-30 minutes

IM haloperidol 10-20mg: 15-30 minutes

IM olanzapine 10mg: 15-45 minutes

IM ketamine 2-4mg/kg: 3-5 minutes


Clearly there is a major advantage with ketamine for achieving rapid sedation where the IM route is necessary.


The registrars discussed how it is important to separate this group of patients from other agitated patients we might encounter such as a disruptive psychiatric patient who might be able to be ‘talked down’ or can be offered oral therapy or elderly patients with agitated delirium who may need IM sedation but without the extreme urgency as a patient who is violently agitated, requiring real-time physical restraint by multiple staff.


The blog discusses 2 pre-hospital studies from Minnesota in 2016 and 2018 which showed a statistically significantly quicker time of onset of sedation for patient given ketamine compared to a control group given haloperidol, as well as a significantly reduced need for re-dosing, which the registrars highlighted as an issue in our practice. If time of onset is too slow, the temptation to re-dose is very high which can often result in over-sedation further down the line. In one of these pre-hospital studies 57% (28/49) patients given ketamine for sedation were intubated, which is thought to be due to the preference of pre-hospital providers transferring an intubated patient rather than a sedated patient without an endotracheal tube.


A 2015 retrospective review of 27 patients given ketamine for severe agitation in the ED, there were few major adverse effects on vital signs and no patients were intubated. An Australian group used 4-5mg/kg IM ketamine to achieve sedation in 49 patients who had already been given sedation without success – this seemed a good rescue sedation protocol for 90% of these patients.


In 2017 the blog author published the first prospective study of ketamine as first-line to sedate acutely agitated patients in the ED, comparing the time of onset to cases where patients received benzodiazepines, haloperidol or a combination of both. Average time to sedation were 6.5 minutes with ketamine, 13 minutes for haloperidol, 15 minutes for midazolam, 18 minutes for lorazepam and 23 minutes for combination of haloperidol + benzodiazepine. A meta-analysis of all studies of patients receiving ketamine for agitation in the ED was 1.8%.


Psychiatric adverse effects – the blog notes that ACEM has a clinical policy that psychiatric illness is an absolute contra-indication to procedural sedation with ketamine; however, a study earlier this year on psychiatric outcomes for patients receiving ketamine sedation in the ED showed no difference in rates of need for psychiatric assessment or admission compared to similar patients given benzodiazepines. A further study is yet to be published on time to adequate behavioural control in the ED for patients given IM ketamine vs IM haloperidol + midazolam.


The blog advises using ketamine as first line for sedation of the acutely agitated patient in the ED – giving 0.5-1.0mg/kg IV/IO or 5mg/kg IM. The trainees recognised the need for care when dosing ketamine for IM injection given the different strengths available – it may be necessary to inject such a high volume that the dose may have to be divided between 2 IM injection sites to ensure appropriate absorption.


Overall this blog provides an interesting insight and a useful round-up of the available literature on the management of what is a stressful situation that requires prompt action and the trainees enjoyed a valuable discussion thereon.

bottom of page