Wessex Intensive Care Society (WICS)

State of the Art Meeting, December 2014, Excel

Author: Dr Emma Fitzgerald
Follow @emma_zcharm6

DAY 1 – Opening Session
Ventilation Trials – Prof A Slutsky, Canada

Began with the background of why this is important.....Review this year in NEJM:
Slutsky 2013 ventilator induce lung injury

A change in philosophy has happened in terms of lung protection.

ARDS Patients

  • The graph in Berlin definition paper (ICM journal) provides a rough schematic on how to ventilate patients
  • ARDSnet 2000 - low tidal volume group showed lower mortality
  • JAMA 2010 systematic review and meta-analysis showed a small decrease in mortality using higher PEEP

Neuromuscular Blocking Agents

Papzian - NMB early on will improve outcome - improved survival at 90 days.

The separation in the kapplein-meyer curves occurred quite late considering that the intervention happened early on.

Prof Slutsky editorial described why this may be the case it may be related to over distension due to dyssynchrony. Seems that cytokines are lower in patients that have been given NMB agent

High Frequency Oscilation

  • non-blinded intention-to-treat MCRCT
  • 548 new-onset, moderate-to-severe ARDS patients
  • HFOV vs low TV high PEEP controlled ventilation strategy
  • outcomes:
    -> 47% vs 35% in-hospital mortality (RR 1.33, 95% CI 1.09 to 1.64)
    -> were given more midazolam, more NMBs, more vasopressors
  • Commentary and criticisms:
    — stopped early due to harm from HFOV
    — HFOV strategy had high mean airway pressures – would a lower mean airway pressure strategy make a difference?
    — groups similar at baseline, both had baseline recruitment manoeuvre to improve lung homogeneity
  • Conclusion: Increased mortality in ARDS patients treated with HFOV

  • non-blinded intention-to-treat MC RCT
  • 795 patients
  • HFOV versus usual care control group
  • outcomes:
    -> all cause mortality at 28 days was 41.7% vs  41.1% (P=0.85 chi-square test)
  • Commentary and criticisms:
    — less hemodynamic compromise, lower airway pressures than OSCILLATE and more protocol variation, possibly due to physician judgement limiting the harm from HFOV settings
    — HFOV groups received more sedatives and muscle relaxants
  • Conclusion: no mortality difference at 1 month

Excellent summary of OSCAR & OSCILLATE from LIFTL here
Commentary above extracted from LITFL


PROSEVA Prone positioning would improve mortality and reduce ventilator induced lung injury. 16% absolute decrease in mortality. This is surprising given the good control arm. They only used PF<150. Protocol at least 16 hours. May need a long time and earlier studies didnt do more than eight hours. The study was performed in very experienced centres.

What about patients with normal lungs.?
  • RCT in CCM (Holland). 2010. Fewer patients developed ARDS with lung protective strategy
  • JAMA 2012.3081651 - consistent with the fact that lung injury was less in low TV ventilation
  • NEJM 2013, 369;428 - patients in theatre. Lower LOS, post op ventilation rate and complications all better in the lung protective group.

  • MV is life saving but VILI has bad consequences
  • ARDS - lung protective strategies
  • Normal lungs patients do better with lung protective strategies

Difficult Decision Making in ICU – Prof P Croskerry, Canada

Many factors that make decision making very difficult. Emergency department throughput and pressures etc are added to in the ICU. There is an increased significance of minor errors.

Diagnostic errors in ICU were investigated. 6000 autopsies. 28% had at least one diagnostic error and 8% of these could have significantly contributed to death. Example used the cognitive reflection test that distinguishes between analytical and intuitive thinking. Published recently in NEJM.

How we make decisions has two effects on patient care? Making the diagnosis influences the treatment. There is an average diagnostic failure rate of 15%. Different specialties are variable. Some have a diagnostic failure rate of up to 50%.

They developed a schematic model to make these work together. Intuition and analytical reasoning go through a process of calibration before making a diagnosis. You can use executive override to influence the output. There is also irrational behaviour that may also change the process. Pattern recognition and repetition can speed things up. Describes a "system 1" (no work required) and "system 2". The expert spends most of their time in system 1 and the novice in system 2. Need to be able to override and toggle between the two.

So, what decision making needs....
  1. Raise the importance decision making
  2. Know operating characteristics of the DOT model
  3. Educate and train intuition
  4. Promote meta cognition and critical thinking

Strategies for Safety – Ms K Vollman, USA

There are pockets of excellence - but variances. For example, 98% of MI patients in USA got a beta blocker. We need a good framework. E.g. structure, process and outcome.

Isn't just about the people, but it is also about the systems that we create. The goal is to reach the error reaching the patient. Process and people factors are both important.

Human factor engineering aims to solve this.

Concept of "silence kills". 80% spotted an error but 58% didn't speak up. "The Silent treatment" 2011.

Suggests use of checklists, huddles and nurse presentation on ward rounds.
Force function is the strongest way of preventing error, e.g. Can't take keys out of the car until you have handbrake on

Industry Symposium - NAVA
NAVA – Dr S Wimbush, UK

IPPV is not very physiological and it causes harm.
Ventilation has moved to spontaneous, awake and less harmful from mandatory, asleep and harmful in the last few years.

NAVA makes the assumption that the vast amount of triggering comes from the diaphragmatic action because of the innervation from the phrenic nerve. Normally we use flow, volume or pressure triggers which means that that the ventilator is always a step behind the patient. The whole aim is to try to restore neuromuscular coupling.

NAVA uses an NG tube that spans the diaphragm. It has an area that detects the electrical activity generated by the diaphragm. It then tells the ventilator to generate a pressure proportional to the signal from the diaphragm. The electrical signal is called the EDI and is proportional to the effort, it is effected by the physiological reserve of the lungs eg a COPD patient would have a different EDI to someone with normal lungs.

Types of asynchrony:
  • Missed breaths
  • Double triggering
  • Flow asynchrony

Asynchrony leads to increased sedation requirements (and the consequences of this), reduced weaning and increased patient discomfort. In the DGH it seems to be useful to get patients off the ventilator more quickly and reduce ICU LOS.

Evidence for NAVA – Dr P Hopkins, UK
Very heterogeneous population
Very little agreement in the definition or description of weaning failure. Also no gold standard in the best mode of ventilation and the strategy for weaning.

Presentation of Kings College data with their use of NAVA.
3 groups....
  1. NAVA
  2. EaDI monitored
  3. Neither

Crude assessment of outcomes. Reduced ventilator days with NAVA.

Suggested using the EaDI monitoring could be useful after extubation. Can help tailor a ventilator and/or airway wean and assess the impact of physical rehab on neural demand.

Suggestion also in palliative circumstances for assessment of dyspnoea when patients are extubated or given a sedation break. Evidence is a bit sketchy, there are largely observational data with small numbers of patients. Recently, a round table has been published in CCM.

They have just started a PS versus NAVA feasibility study. Intervention group uses EDI and NAVA at night, as you would with a PS wean. I.e. they are testing the potential benefit for the neural trigger and also for the use as a monitoring device to prevent OVER ventilation and OVER sedation.

The Cauldron

Welcome to the uncaring profession – Dr Adrian Wong
Adrian begins with some philosophical reflections about medicine in general. He emphasises the value of a ‘good’ death and the fact that long term outcomes are increasingly used to assign value to ICM. There followed advocacy of guidelines, checklists and protocols.

What is the difference between care and therapy?? Emotions influence care and they cause variability. Being human is a weakness. Thinking isn't for everyone. We need fewer strategists and more doers!

I think, therefore I am - Dr D Garry

The intensive care units of the future are the ICUs of the mind.....
Since the 1950s the ICUs have focused on treatment of PHYSICAL problems.

How do we trust our senses?? Cogito ergo sum....I think therefore I am. Plato - the treatment of the body and the soul should not be separated.

We need to aim to provide the brain with an environment whereby it can think clearly. We need to separate it from the damage caused by a critical illness.

The intensive care units of the future are the ICUs FOR the mind..... If I am not thinking, I am not.

The price of predictive modelling: eliminating bias by eliminating doctors from the decision making process - Dr A Devansen

All humans are prone to cognitive bias. Predictive modelling is a scientific theory to provide accurate and predictive information. Modelling in practise is subject to user bias. Models perform better than people.

"The world of economics is light years ahead of medicine as they use mathematic models and also have no scruples!"

Human fallibility. Machine perfectibility. Is the hospital of the future run by machines?? You would have to have OBJECTIVE inputs to the scoring systems that do not have any objective measurements, e.g. the Wells score would. It work as you have to put a clinical opinion into it that uses someone's clinical judgement about whether a PE is or isn't likely.

The geriatric intensive care unit - Dr Eryl Davies

There are lots of similarities between ICU patients and geriatric patients, e.g. physiological extremes, poly pharmacy issues and end of life discussions.

There may be an influence of the recent publicity of the LCP in the referral of elderly patients to the ICU. Is it appropriate to offer "heroics" in the elderly population? Is age just a number??

Creation of a dedicated geriatric ICU was not a popular option amongst ICM consultants.

Elderly patients are complex and need specialist intervention, knowledge and care.
Closing statement....geriatric ICUs are the future.

Intensive Care's Weapon of Mass Destruction: The Random Misā€Controlled Trial - Dr P Nandhabalan

Started by looking at HFO trials and the other studies that have contract iced each other. begs the question..is it the INTERVENTION or HOW we assess efficacy??

Why are RCTs so dangerous?
  • Heterogeneity in many factor and so the ICU population remains very poorly defined.
  • Design - exclusion and inclusion criteria are often strict, regional differences.
  • Outcome measures are arbitrary
  • Changing disease patterns
  • Statistical lies!!

Looked at steroids, HFO, tight glycaemic control and vasopressin.

The heroes and villains often switch around.
a weapon able to cause widespread devastation and loss of life... This is ONE occasion where further large scale RCTS are not required!!!!

Day 2 - Trauma
Trauma and coagulopathy – Dr K Brohi, UK

Loss of fibrinogen is one of the biggest problems in traumatic bleeding. Cryoprecipitate is the best product to give, but we tend to not give a big enough dose to replace the fibrinogen and restore levels to normal.

Cryostat study - a pilot study
Preliminary results were presented. Early cryoprecipitate seems to reduce mortality, BUT the groups were small and there were significant differences between the groups. They are preparing for a bigger trial.

Fibrinolysis is also a problem. ROTEM doesn't seem to pick it up.

So, trauma patients are in a state of fibrinolysis and also have low fibrinogen levels. If you can modify the production of activated protein C, you can abolish traumatic coagulopathy (in mice). Trauma patients release tPa, which then triggers APC. Therefore tranexamic acid is important even if the ROTEM doesn't suggest it is needed.

Need to focus on the fibrinogen replacement and avoid dilution with crystalloids.

Trauma induced lung injury - Prof S Jaber, France

Many causes of ARDS post trauma...primary and secondary (direct and indirect) causes.
CXR, echo and CT can all be used to identify causes of lung injury.
Recent study used volumetric software to evaluate the degree of lung contusion. ROC curve showed that if you have 21% loss of lung volume, you have an acute lung injury.

Trauma patients are at very high risk of developing a lung injury. They have many risk factors.

The impact of it - increase in mortality if a trauma patient develops a lung injury (OR 1.6)

Prevention is an important way to improve the outcome...
  • Limit fluid overload
  • Limit transfusion
  • Use NIV where possible
  • Lung protective ventilation

Article in American journal of respiratory and critical care medicine.... implementation of evidence-based extubation readiness bundle in 499 brain injured patients.

Acute Kidney Injury – Anything New?
Urine output and Urinalysis – Dr R Beale, UK

Urine output depends on the scenario.... Solute load and concentrating ability

Kidneys filter about 1.5kg of salt per day. Maximum concentrating ability is 1400mosmol. Solute load is 700mosmol.

The pre-AKI is possibly the intervention phase, once AKI is established it is probably too late to improve things.

Physiological oliguria.... RAAS, ADH, sympathetic nervous system. Oliguria may imply a physiological or a pathological response.

In terms of differential diagnosis, oliguria adds very little. It can be of a prognostic use...AKI with oliguria may correlate with a poorer outcome.

Anuria - probably the most important finding. Is obstruction or lack of peer fusion until proven otherwise.

Modified urine output.... Using frusemide.
  • No evidence that diuretics can alter outcome, but there is a signal that if people are passing urine then they have a better outcome compared to those who are not.
  • CCM paper this year...frusemide stress test to predict severity of AKI. IV frusemide was given. Patients that responded to frusemide had a higher rate of recovery. Probably tells us that the tubules are still working (the frusemide needs to be transported across the tubules).

Urinary electrolytes
  • Fractional excretion of sodium test. Is supposed to differentiate between pre-renal failure and ATN. Many exceptions may it pretty useless in the ICU e.g. timing. CCM paper (Darmon) dismisses it.
  • Fractional excretion of urea is similarly useless...

Can sometimes be useful, e.g. GN (casts), porphyria, rhabdomyolysis

  1. Urinalysis isn't that helpful
  2. Volume balance is important

Biomarkers for AKI – Dr J Prowle, UK

We are unable to risk stratify patients as we can't recognise AKI early and the treatment or intervention we need to do needs to be done early for it to be effective. Therefore, we need to develop better predictive markers of AKI.

Ideal marker would
  • Enable risk stratification for the likelihood of AKI but is biochemical AKI the right measurement?
  • Prognosticate
  • Provide the earliest possible diagnosis
  • Reliably distinguish tubular and pre renal causes
  • Identify sub-clinical injury if it exists

Creatinine is a poor marker of renal impairment. We need better biomarkers. Some are in the blood, some in the urine. Some are established e.g. Cystatin 3 and beta2microglobulin. Other novel ones such as NGAL

  • Seems to be significantly upregulated in the kidney in ischaemic conditions.
  • Source of NGAL - produced in the distal nephron.
  • The AUC was good in Paeds cardiac patients (0.98) but is only 0.7 in ICU patients. Is not very useful in chronic kidney disease.
  • Urinary NGAL may be useful as a triage tool in ED.
  • What about high NGAL and normal creatinine? May have sub clinical AKI. Not really sure what the significance of this is.

Procalitonin levels have been shown to be related to development of AKI.

Kashani, in critical care 2013 17-R25 identifies 2 biomarkers that could predict the development of AKI, even multiplied them together to improve the detection. AUC was above 0.8. Choice if threshold is difficult. Likely to be marketed with the choice of two thresholds. These biomarkers are associated with G1cell cycle arrest and appear therefore to be linked to the development of cellular injury.

What do we do with biomarkers results?? We probably need to gather more evidence. the "treatment" is largely supportive care and preventing secondary injury, but we do this anyway... Maybe better to use them to triage or de-escalate care.

How best to prolong extra-corporeal circuit lifespan – Dr M Ostermann, UK

Why does the filter clot?
  1. Haemoconcentration
  2. Thrombogenic clot surface
  3. Exposure of blood to air (ie in drip chamber)
  4. Risk factors (e.g. SIRS, filter size, hypercoagulable patient, lipid infusions - some are modifiable)

We could manipulate the following...
Pre-dilution associated with a longer circuit life, at expense of reduced clearance
Membranes: No good RCTs comparing different membranes with regards to circuit life
Choice of modality: Suggestion that CVVHF is associated with more clotting, but poor evidence
Heparin: LMWH vs heparin - no evidence
- Weak anticoagulant
- Limited number of studies. Alone, it is inferior to it being combined with heparin, but is an alternative to people who can't have heparin
Others (eg argatroban, anti thrombin, fondaparinux): No real evidence
Regional anti coagulation - Citrate, heparin/protamine

Citrate: More popular than heparin/protamine for regional anticoagulation
- Binds calcium, thereby inhibiting the clotting cascade at multiple steps. Calcium needs to be added back to the blood as it goes back to the patient. Citrate is eliminated as the blood flows through the filter. About 50% of the citrate goes back into patient
s systemic circulation, slots into Krebs cycle and gets metabolised to bicarbonate, can be quite useful in patients with renal failure
- Does it work to keep the membranes patent? Six RCTs have been done
. Largest one = crit care medicine. 2009(37):545, Helene et al. No change in filter lifespan. Equally efficacious, but citrate was safer.
Two meta-analyses have been done
- Wu et al Am J Kid disease 2012;59. Some RCTs show increased circuit survival... Suggested a trend in the direction of citrate, but undoubtedly, citrate was safer in terms of bleeding risk.
- Different analysis by Zhang et al 2012 ICM. They differed in final conclusion about lifespan but agreed citrate was safer.

What should we do in real life?
2012 kidney international (KFIGO) 2;1-14. Felt there was enough evidence that in patients receiving CONTINUOUS RRT, citrate should be first line, heparin reserved for those intolerant of citrate.

The RCT results don't always reflect the real clinical goings on. Data from the first year of using citrate at guys and Tommy's was presented. (Towey et al nephron clin practice 2013: 124.119-123):
  • Citrate performed better than heparin and prostacyclin
  • Rate of premature clotting was reduced by 76.5% with citrate
  • No difference between heparin and prostacyclin

  1. Anticoagulation is not a substitute for poor access
  2. Choice should be influenced by local expertise and patient characteristics
  3. KDIGO suggest citrate as first line anticoagulant
  4. Citrate offers regional anticoagulation with less bleed don't and potentially better circuit latency
  5. Protocol and attention to detail are essential (independent of type of anticoagulation)

Trainee Research Session
Preparing and writing for publication

Understanding the publishing process
Write first for:
  1. Editors
  2. Reviewers
  3. Your audience

What drives editors?
Which papers get cited most (long term)?
  1. Methodological papers
  2. Discussion papers and concept analysis
  3. Review papers
  4. Original research i.e. the further down the list, the more it has to shine

Which papers get cited most (short term)
  1. Review
  2. Methodological
  3. Discussion papers including concept analysis
  4. Original research

Process of writing - 4 rules of writing
  1. Read the guidelines
  2. Set realistic targets and count words
  3. Seek criticism, don't write in isolation.
  4. Treat a rejection as the start of the next submission

Journal guidelines
  • Look at the guidelines and write to them with ideas and find out how they want it to be written
  • Length... Gives an idea about dividing the contents
  • Layout... Font and spacing, wide margins.
  • Organisation of the paper... Takes you form the broad idea, to discussion to conclusion. Use the headings that they suggest. Think about writing the paper before you have done the experiments. Don't start at the intro, start with the method and work backwards as you don't know what you are introducing yet
  • Other conventions
- Spelling (esp North America)
Presenting stats
- Number (words, numerals etc)

Setting Targets: All good writers do this. Try to write a specific number of words per day or every time you sit down to write. When you have reached your target and STOP....!!

Seek Criticism - Find a "critical" friend

Understand the peer review process: Not a scientific process or black and white, about judgement
Follow international guidelines:
    Creating, Presenting, designing and promoting a poster
    No point in doing research unless you deliver your message.
    Design courses are very useful. They are very personal too though.
    Poster design
    • Content
    • Anatomy and circulation
    • Rules

    Communicates visually

    Attracts and holds attention
    Concise and organised
    Stands alone when you are not there

    Needs to be readable from 6 feet away.
    Needs to deliver the message in ten seconds
    What do you want to add to this field of research.
    What are your conclusions?
    What are your messages?

    40-50% text
    >30-40% graphics
    <20% white space

    Title... Think really big. Convey the theme, needs to be catchy max 1-2 lines.
    Contact details
    Intro - should place your work into context.
    Aim - 1-2. Lines. What am I adding?
    Methods - important, but limit. How am I adding knowledge? Bare essentials, not to the details of the manuscript. Flow-charts and diagrams are useful. Adding links may be useful.
    Results - always the largest section.. What did I find? Graphs are good. Figures better than tables
    Conclusion - give your interpretation of the data.
    Take home message

    Don't fight reader gravity. Down and then across.

    Clarity: Attract, Inform, Engage, Conclude
    Title with a message
    Make findings obvious
    Link images with the text

    Use graphs well
    Take home message and killer bundles

    PowerPoint most popular.
    Adobe illustrator is more difficult to learn, bit good design software.

    Use sans serif fonts. Calibri or arial
    Keep to 1/2 front types
    Avoid excessive text and out breathing space around your words
    Font size - should be able to read the smallest font from 6 feet. Ascribe level of importance through font size
    . Tittle 85pt / Authors 60 / Subheadings 36-54 / Body 32

    Colour and contrast
    Light background with dark font is the easiest to read. No more than two or three colours. Not blue on red, or yellow on white or red on blue

    Communicate concepts quickly.
    Not too many
    Ideally should see from six feet
    Avoid 3D and hashed ones
    Get rid if grid lines and backgrounds so that you can convey tour message
    Resolution at least 150dpi, ideally 300dpi but no higher

    Other things to consider
    Add-ons: QR codes to link supplemental materials, contact emails, the poster, your website
    Screen versus print. Bear in mind the back lit monitor as you would with photos

    Design and layout summary
    Justify to left
    White space and balance.
    Font sizes and colours

    Types of poster sessions
    Thematic poster - related abstracts under a few main themes, discussion with moderator
    Poster discussion/corner/e-co. Session/ViPER - computer session. Ecomm max 3 slides & 5 mins
    Imagine your stage and prepare.Who is my audience?
    • Body language is important
    • Check and show your poster
    • Prepare a 3-5 min presentation.. Focus on discussion
    • When you are there....attract, engage, inform, conclude. Sound excited, even if you are the only one!!! Get the message across, know your limitations and conclude. Ask for their contact details. Give a summary sheet or have a QR code.

    The future... Video links, screens not paper etc.

    Organ Donation
    Task force says that we have to consider that organ donation should be a part of EOL care and that if we put donation leads into our hospitals and we get the infrastructure we need, we should get more people having donation as part of mire people's end of life care.

    GMC 2010 said doctors have a duty to offer donation, NICE reminded us of when donors should be referred.

    Government accepted all of the recommendations from the taskforce. Baseline was 2007. Have seen an increase in donation. 49.7% increase in donation. Increased DCD. More transplants done than ever before. The waiting list has also gone down.

    Progress continues. This year, until dec 8th compared to previous year, the numbers have gone up by a further 13%. The transplant rate has also gone up from 2000 to 2500.

    The new target is to match the performance of world class donation and transplant programmes.

    The problem with hitting a target, is being given a new target. By doing more brainstem testing by making it more a part of EOL care, will we refer more patients as there have been many who met the criteria for BSD testing, that were not tested.

    Retrieval and transplantation need to have the same level of implementation as critical care. W must maximise the gift of donation. Is regional reperfusion or explant and recondition the organs. Is me it about variability of decision making?? One unit turned down 80% of kidneys offered too them, despite them seemingly being "ok".

    However, we will never have a fully functioning system whilst 43% of the families are saying no.

    Society still views donation as the exception, not the norm. Defaulting to "no" seems to be the best option as people feel more comfortable not donating organs in someone who would have wanted to, rather than donating organs from someone who wouldn't have wanted to. Maybe we need more hardcore advertising programmes... I.e. "By donating you are saving Jill, by not you are killing jack" idea.

    Discussion around the concept of introducing organ donation consent when having a procedure or operation. Controversial as patients may think that it will influence their subsequent care maybe though it should be discussed even more in other situations.

    Discussion about the concept of donating kidneys in the patient before the patient dies. They won't die from renal failure and it would reduce the warm ischaemic time. Not taken well by the panel...thoughts were to "sort out society" first and mend/establish a good system for dead donors before tackling this and the problems it may/will cause.

    Reference...BJA 2008 June editorial.

    Pot Pourri of Infections
    Optimising antibiotic therapy and the DALI study

    Focus of antibiotic therapy is largely still all about timing.
    The dose - one size, does it really fit all?? We are told to give an "appropriate" dose...

    Antibiotic dosing is related to
    • The microorganism - SENSITIVITY
    • High MIC, but still susceptible bug, have a high chance of failing treatment, eg pseudomonas and MRSA. Don't always get the MIC data from the lab, there is quite some susceptibility of the microorganism to the antibiotics you are giving.
    • Volume of distribution and pharmacokinetics change when patients are critically unwell as compared to healthy fit and well patients. Most of the drugs - there is an increased VD to 2 or 3 fold. Most of the time we are only concerned when the clearance is low, however the opposite is more often actually a problem. Hyper filtration causes increased removal if the drug than you would probably want. Calculated cr clearance is the best way to detect this problem. The other drugs and fluids we give will increase the clearance
    • Incidence of ARC (augmented renal clearance) was high in 52% of patients, 18% permanently (Ghent data). Was related to drug (antibiotic) failure...High risk patients include low sofa score, burns patients.
    • The drug and it's PHARMACODYNAMICS.
    • Giving a dose results in a concentration. How this concentration results in a clinical effect is the pharmacodynamics. Some drugs are time dependent (eg b lactam) need a prolonged exposure, amino glycosides need a high peak as they are concentration dependent. We can use this to our advantage by adjusting the dose to ensure the efficacy with minimal toxicity.

    Potential consequences of under dosing
    • treatment failure
    • need for multiple antibiotic courses
    • antibiotic resistance

    DALI study - Primary aim was to describe the ok of the beta lactam antibiotics point prevalence. International study. Blood samples taken. Slightly different timing for different antibiotics and those on continuous infusions. Used the EUCAST MIC values. They investigated whether the patients actually reached the required antibiotic level. The sample population was relatively representative of a typical ICU. Overall mortality 24%.

    1. clearance for the drug was lower
    2. VD - lots of variation, but generally higher
    3. Enormous variability in antibiotic concentration.
    4. Confirms that that the concept of "one size fits all" does not ring true.
    5. Minimum target - only 80% reached the minimal beta lactam target...20% not getting an effective dose

    Drug exposure and outcome....the time above the MIC was correlating well with the success of treatment. Ie probability of cure is related to the drug level.

    Factors associated with reaching the PKPD targets...only 2 were associated with reaching target
    • creatinine clearance was negatively correlated with reaching the target, ie augmented clearance was bad.
    • Use of continuous infusions - 4x increase in chance of achieving the PKPD target

    How to solve this?
    1. Everybody gets double The dose, but you may cause toxicity
    2. Front-loading - extended and continuous infusion with a loading dose, or a double first dose if dosing intermittently.
    3. The use of continuous infusion...if you increase the infusion time, you increase the PKPD, i.e give the same dose over a longer period of time, rather than having a peak and a rapid decline.
    4. Continuous infusion of meropenem gave better outcomes in a feasibility study and a better clinical outcomes as well as a better PKPD

    "We can and should do better. We should use the antibiotics we have in a better way"

    There is no single patient characteristic that improvement things, but augmented clearance seems to be important..

    Why aren't we doing selective decontamination?

    Hospital acquired infections seem to come from the organisms that are colonising the patient. Enteric gram negatives and yeasts. Doesn't target anaerobes

    Topical and enteral treatment. Short course of IV antibiotics.

    Does it work? Cochrane review
    • Reduction in rates of pneumonia and improve survival with topical and IV
    • Reduction in pneumonia but not mortality with just topical
    • Cluster RCT was excluded. De Smet 2009 NEJM
    - Crude mortality not different
    - Had to adjust for the differences caused by the cluster randomisation

    Not used in the UK - not enough evidence and there is concern about antibiotic resistance. NICE specialist advisors commented that there were concerns about C. difficile.

    Cuthbertson CCM 2013 - Clinical stakeholders' opinions of SDD
    Broke it down into their knowledge, professional role and their beliefs
    Most people were not strongly opposed, but were uncertain
    Skew towards a belief it causes antibiotic resistance
    People thought that it hadn't been adequately addressed....so they did a SR (lancet this year) for all the point estimate favours SDD, except for MRSA which is above 1

    De smet paper ended up giving less days of broad spectrum antibiotics in the SDD group overall

    Is it generalisable?? It does seem to be a concern.... Lots of evidence comes from the Netherlands. But there is some evidence from the UK and USA.

    UK evidence
    • B Winter. Small RCT 20 years ago with historical controls. Only 90 patients. Concluded it could be safely used in the ICU.
    • Cheating cross experience 2007 - SDD introduced in 2007. Reduced rates of MRAB. MRSA certainly hasn't increased, VRE not really gone up.
    • Email survey (unpublished) 13 units said they use SDD. ICNARC looked at the outcomes from these units. Antibiotic resistance does not seem to be increased. Mortality is adjusted for units and there doesn't appear to be much difference. Caution comparing the different units. In those that used the IV component had a reduction in blood stream infections
    • SuDDICU
    - people equivocal about c diff. No data to say it does increase it
    lots of uncertainty about the benefits.

    SDD - all of the consultants need to be involved. Can't just have some people doing this. Intensivists seemed to be less opposed than the microbiologists

    Direct quotes HPA report
    • Data from Netherlands doesn't count
    • We just never discussed it
    • Haven't gotten rd to it yet
    • Still worried about resistance, but I know the evidence doesn't Support this
    • Need an SDD champion

    1. very few strongly opposed
    2. lots of equipoise
    3. 85% were keen to be involved in a big international RCT with a hard endpoint and monitoring of antibiotic resistance during and for a prolonged period afterwards
    4. new large cluster trial is being designed. Resistance monitored for a year afterwards no cross over as they want to see the ecology
    5. Need 24000 patients

    Experience of a novel (beta) corona virus infection
    June 2012 HCoV-EMC. Closely linked to bat virus. First human case that made it to the ICU NEJM case report
    Case report of a patient cannulated for vv ECMO for refractory hypoxaemia

    Day 3

    Review of the Ely study from the 1990s.
    Study was repeated and published in the blue journal in 1994 - showed no difference

    Protocols may be useful, but probably not in the difficult prolonged weaning phase patients.

    Presented the data from their unpublished study.

    6% of patients in icus in Scotland require a prolonged period of mechanical Ventialtion. 5-10% seems to be the figure that represents the number of patients facing a prolonged weaning period.

    45% of the weaning failures manage to get off the ventilator. 20% get to NIV. Neuromuscular disease patients have the lowest mortality, COPD patients have a shorter wean but the highest mortality.

    We need to reverse the reversible.

    Description of the MDT approach to the management of the weaning patient. Described...therapy, physiology, exercise etc...

    Concluded that we need bespoke weaning and rehab plans for these patients. We also need to ventilate them properly at night.


    Has become a fast moving speciality in the last few years, after the creation of stroke units and hyper-acute services - hyper acute stroke units....

    Stroke units have reduced mortality. Thrombolysis has improved functional outcomes. If you bleed though, there is a 70% mortality at one year.

    The data for strokes is quite old and probably doesn't represent the current improved practice in the UK.

    Overall mortality in the uk in Icu is about 39%.
    OOHCA 50%
    Oscillate mortality around 30%
    BMT 69%.

    Ischaemic cva patients...if they survived to discharge... 66% chance of favourable outcome, 33% gain independence.

    Description of the development of the surviving sepsis campaign and the guidelines.

    Controversies and changing evidence led to the reviewing of the guidelines. There has been a guide to the guidelines written by JL Vincent and J Marshall after the publication of the 2008 guidelines. There were then a series of papers published to look at the impact of and adherence to the recommended guidelines. It had improved but was not good enough.

    The latest guidelines change the bundles. The management one was dropped and the resuscitation one was split into two phases. We are in the "reinvigoration phase" of the surviving sepsis campaign. There has been a suggestion that they should be mandatory which clearly has big issues.

    There has no been a moved to completely avoid the involvement of industry in the creation and implementation of the SSC. More involvement from professional bodies, but strong feeling that industry involvement is bad.

    For better or for worse, it seems that the bundles that we create are increasingly being used as a marker of performance.

    The positive effect of bundles was more marked in severe sepsis rather than in septic shock, suggesting that there is perhaps more opportunity to intervene in these patients. Most of the things that make a difference are PROCESS things.

    The new guidelines dropped a lot of controversial, harmful and out of date things. A lot of these actually are things that are too late, ie the horse has already bolted. The bundle approach has been successful in achieving a reduction in mortality, whether it is because it is an operational lever or due to the individual components is unclear... But does it really matter???!

    Charles sprung lecture...does one size fit all?

    Should the CVP target be changed in patients with heart failure?
    Should a MAP target of 65 be higher in those patients with hypertension?
    A BP target should be titrated to your patient. 

    Is the urine target the right one in patients with renal failure?

    Should you change your targets in those patients being fed? Or in patients with diabetes?
    Swings in glucose are probably bad

    Typically using 6ml/kg. 
    There is some data to suggest patients with normal compliance with low TV had higher mortality

    Most people agree that an Hb trigger of 7g/dl is what we should use

    Hard to define adequate fluid resuscitation and vasopressor limit to trigger steroid administration 

    The slide presentations for the data giving the guidelines by each author are almost all completed now and are published in Critical care med 2013 (41) and ICM 2013 (39). 

    Mervyn Singer Lecture
    One size cannot fit all. The simple rigid protocol may be applicable in a small closed community, but we should be using generic guidelines instead that we can adjust to our specific patients.

    Initial empirical antibiotic therapy... Challenges the benefits of early antibiotics. Critical care med 2011.
    Other papers also suggest that maybe taking cultures and waiting to give antibiotics. There was a lower mortality in the less aggressive group. Lancet infection. University of Virginia.

    Antibiotics may not be the panacea that we think they are.

    ICM journal last week... De-escalation of empirical therapy is associated with lower mortality in patients with septic shock.

    New York have started using them as protocols NOT guidelines, meaning that patients are being given MINIMUM amount of 30ml/kg I fluid. Looks like the sepsis bundle will be a standard of care across America. Massive implications.

    "Benchmarking the incidence of severe sepsis in the US"...
    Coding is becoming important. How are we diagnosing it?? However, promos and Process have really struggled with recruitment.

    1. Principle is great
    2. Need to be cautious and embrace the guidelines but we need to combine it with individualised expertise

    Resuscitation End-points
    • Remember the balance between production and clearance. Sometimes is very difficult to distinguish between some of the conditions
    • Can you use lactate to guide resuscitation? The lactate pyruvate ratio may be important. The interventions should focus on he early stages when tissue hypoxia is occurring
    • More rapid decreases in lactate are related with better outcomes, but is it the manipulation of the lactate that is responsible?
    • Changes in lactate seem to be important in identifying the patient that may benefit from your haemodynamic interventions.
    • Not much data to suggest whether targeting lactate levels is beneficial. Target to 10% is likely not enough

    Lactate guided therapy study in the blue journal.

    Maybe lactate clearance is important. Need to reduce production. If you bind and remove the lactate and try to remove it, you worsen hypoxia especially in the heart. Need to stop the fire.

    Can echo be used to titrate therapy during resuscitation?

    Echo can help to establish a diagnosis and avoid unnecessary interventions.

    • IVC diameter. Above 2cm the pt is unlikely to respond to a fluid challenge
    • Changes in diameter during spontaneous respiration. Complete collapse seems to be related to fluid depletion. In positive pressure ventilation you can use the distensibility index
    • Measurement of contractility. "Eyeballing"
    • LVEDArea. Gives you information quickly. You can use the fractional area change
    • Remember to consider the whole picture

    Not much evidence out there but is being used a lot more. There is a BJA article about the impact of echo.

    Inter-observer variability
    Discontinuous measurement
    Measurements are affected by angulation
    Poor views affect results.

    Closing Session
    Management of therapeutic hypothermia (TH) after cardiac arrest. Prof Kjetil Sunde, Norway

    After successful CPR, normally the temperature increases within the first 48hours. This is worse in those who have the worst outcomes. The post cardiac arrest syndroms have 7 stages. It is a severe repercussion syndrome associated with haemodynamic instability.

    TH inhibits several aspects of the reperfusion. TH reduces cerebral blood flow and so there is a 10% reduction in oxygen consumption per degree in temp drop.

    Initially non randomised trials and then 3 RCTs showed a benefit from TH. Led to the ILCOR statement suggesting cooling post cardiac arrest to counterculture the reperfusion syndrome which appears to be unrelated to the rhythm causing the problem.

    Most studies since have shown improvement in outcomes and certainly no harm was detected.

    The whole point seems to be the reperfusion syndrome.

    BUT we don't know the therapeutic window, the ideal temperature nor the duration.

    NEJM TTM trial was discussed
    • Led from Lund in Sweden. 950 patients 10 countries in Europe and Australia. 36 hospitals
    • Prognostication was blinded and outcome was blinded
    • Followed for 6months after discharge
    • No major differences between groups. High bystander CPR time to ROSc was about 25mins

    No major differences in survival at any point in he trial
    Excellent follow up
    CPC score - no differences

    Adverse events (in the supplement)
    No major differences but more low k in cooled group and tendency for bleeding.

    Targeting 33 - no benefit.
    Excellent study
    .Strict criteria and prognostication and excellent control.

    Discussion points
    1. Only one minute to starting CPR. Therefore very low anoxia time. Although it was pointed out later that the time to ROSC was still quite long. Suggestion from other papers (resuscitation journal) that cooling IS beneficial if the downtime was more than 3 minutes, thereby making TTM trial difficult to interpret and compare cooling benefits
    2. The 33 degree group should have been prognosticators later due to the effect of cooling on sedation. They might have lost survivors in the cooled group. Suggestion that there are a good number of late wakers after cooling. These could have been CPC 1 survivors. We measure good outcomes with CPC1-2. For the relatives the difference between these is actually very important. We should aim for good CPC 1s.
    3. The period from discharge to 6m. Significantly more cooled patients got up to CPC 1 from2 than in the other group. What about the following 6m/12m etc? May again be missing people.
    4. In animal studies, there is a suggestion that time to target temp doesn't matter. Some clinical trials suggest it maybe matters. JAMA this month - 2l saline to cool pre-hospital didn't improve outcome

    Problems with cooling
    • Shivering - need a sedation protocol
    • BM disturbance - insulin resistance. Need insulin infusion
    • Hyperventilation and low co2 can be a problem.
    • Haemodynamic instability can also be a problem. Need to maintain vital organ blood flow. Can have issues with bradycardia, patients with a lower HR seem to have a better outcome at any point in the cooling process. Don't treat it if the Bp and lactate are ok

    Use targeted temp management
    Avoid fever
    Suggestion that we need more data and we shouldn't withdraw care too early
    Advised to await ILCOR and probably continue cooling for now

    ILCOR have since released an interim advisory statement

    Acute pancreatitis. Professor Jan De Waele, Belgium

    New scoring systems

    Updated Atlanta classification
    Modified Marshall score

    nfection and organ dysfunction are the two drivers for a bad outcome. Essentially need a CT for Modified Marshall score as you need to identify necrosis. Not needed for Atlanta score
    Previously used biochemical, physiological and radiological scores. These scores seem equivalent. Concluded that you don't need radiology to confirm diagnosis nor to prognosticate
    BUN may be useful. Lots of variation though
    Fluid over-resuscitation is likely to affect the micro-circulation and this can not only affect organs, but worsen the pancreatic necrosis. Not giving enough fluid equally is not good! Trial in Chinese medical journal - more restrictive arm did better on many counts
    Intra abdo hypertension is linked to organ dysfunction. Percutaneous drainage seemed to help and improve outcomes. Decompression may be needed but not much evidence. Usually last resort.

    Some other useful resources from the meeting

    Blog from the ICS

    References from many of the presentations available

    Storify feed from ICS meeting available

    Summary of presentations attended by Dr J McNicholas (consultant at QA Hospital, Portsmouth)

    blog comments powered by Disqus