Blog

Todd Fraser from Australia wrote 05-31-2013 07:49:14 am
Well, the 2013 CICM Annual Scientific Meeting is underway.

This year’s conference, entitled “Down With Dogma”, is setting out to ‘challenge the fundamentals of Intensive Care’. The first CICM ASM on the eastern side of “the ditch” with College President Ross Freebairn from Hawke’s Bay, and Wellington’s own Paul Young refereeing session number one : Time for some radical thinking.

Young begins with the origins of the meeting – challenging the suggested theme, he raised “Down with Dogma” as an alternative. Proving that having good ideas was bad for you, he suddenly found himself as the convenor.

Dogma, he says, is the foundation of medicine – we learn by watching what others have shown us, accepting its truth on face value. ICU may be even more vulnerable, as the instant gratification provided by monitoring may reinforce our belief that what we’ve just done is good for the patient – and we’re forgetting to take a long term view.

Todd Fraser from Australia wrote 05-31-2013 07:49:31 am
The session begins with one of the international luminaries, Prof Derek Angus, who quite reasonably asks when will get that magic bullets in sepsis are just not going to happen.

Sepsis remains the “disease of ICU” and despite decades of well intended research, a major breakthrough in sepsis management remains elusive.

So how do we address this? Many limitations exist – heterogeneity of patients, getting appropriate sample sizing, poor background evidence in laboratory and animal studies, the progressive reduction in mortality (making mortality outcome trials difficult) and a reluctance generally to enrol patients.

Currently there are about 500 sepsis trials available in the literature. Angus says that nearly all of these appear very similar to trial design that has gone before – perhaps this is the very Einstein definition of insanity…

One of the prevailing dogmas of sepsis remains that the most important factor is the bug. Not so, he says, it’s the patient that is the problem. While effective antibiotics are important, clearly many patients die despite this. The pro- and later anti-inflammatory responses resulting from severe sepsis clearly have much to do with it. The process remains poorly understood, and is known to be incredibly complex – how is a study of a single agent to be expected to impact on this?

So why do some patients get severe sepsis, and others do not? Different cytokine profiles can be demonstrated in mild sepsis, severe sepsis survivors and those who died.

Sepsis then reflects a balance between pathogenicity, host response-related damage, and the host tolerance to each of these.

Some of the mechanisms of these pathways relate to our genetic makeup. Specific toll like receptor differences provide resistance to some pathogens and exposure to others – and these can be impacted upon by environmental factors – for example one is selected for in Subsaharan Africa where it protects against Malaria, but is associated with reduced resistance to gram negative sepsis.

The trick now will be identifying these genetic differences in practice, then being able to do something about it. To do so, we need to establish enough knowledge of clinical, genetic and molecular factors associated with sepsis – this is extremely under-developed at this point, but baby steps are being taken. This will lead (hopefully) to drug target generation, biomarkers that identify drug responsiveness (examples including corticoid resistance in steroid therapy) for specific patients and biomarkers that guide ongoing therapy.

So how do we overcome these limits? Adaptive trial designs may provide some hope (to be discussed later). In these trials, ongoing monitoring of biomarkers in the trial can enable you to identify the ones that predict effect, and the trial is gradually morphed to look specifically at biomarker-driven randomisation.

Todd Fraser from Australia wrote 05-31-2013 07:49:56 am
Next up is Auckland ICM giant Stephen Streat, with ‘Cults, Memes, Gizmo Idolatry, Wiki-truth and the rise of Magical Thinking’. There’s going to be some straight shooting here.

Streat promises up front to disturb our “conference comfort”, and I’m sure he will!

Streat suggests that much of what we know to be true is influenced by what we desire to be true. He cites tight glycaemic control as an example par excellence. How did we become so distracted by this without an appropriate level of understanding?

He continues, quoting DECRA and SAFE as an example of clinicians interpreting the data differently based on non-scientific factors.

Wiki truth – truth can be edited to support the beliefs of the user. As Wikipedia itself says, truth and verifiability (ie, that it has been written elsewhere), are not the same thing.

Now to technology, he suggests that we are guided by the beauty of our weapons (obviously a Leonard Cohen fan) – the rise of Gizmo Idolatry. We intensivists are obsessed with our gadgets and believe in them with religious fanatacism, to the detriment of our patients.

He concludes with a call to arms to act responsibly and ethically in our assessment of the information we have at our disposal.

Todd Fraser from Australia wrote 05-31-2013 07:50:34 am
Next up, Charles Natanson, a veteran of over 30 years as senior investigator with the NIH in the US, challenges us with his talk – “how practice misalignment creates the illusion of knowledge”

This is a fascinating area of conflict in critical care research. In an attempt to provide a standardised control group, therapies that are usually titrated to the patient’s condition are often prescribed in a fixed dose. This results in the inadvertent creation of subgroups within the control group that are over and under dosed compared with the real world, and may artificially worsen outcomes in the control arm.

In addition, the intervention arm will also be affected, with some patients being inappropriately over or under treated.

In essence, the randomisation process may result in patients with severe disease getting minimal care, and vice versa.

The trial outcome then becomes a choice of the “least-worst option”. He quotes two pillars of our practice – TRICC and ARDS-NET low volume ventilation as examples.

TRICC, he says, reinforces this. The sentinel trial of restrictive transfusion practice in ICU randomised patients to transfusion at lowish or higher haemaglobins. The control arm, used as a comparator on the basis that it is “usual practice” does not actually reflect this – a survey of Canadian intensivists at the time indicated that some of the patients in this group would have not have been transfused. Thus, some relatively well, young patients received tranfusion, and relatively sick, elderly patients had transfusion withheld. As a result, he says, we have exposed patients in both groups to harm, and the trial result reflects the balance of these two harms, and we lose the opportunity to compare a new approach with “standard practice” – this affects our understanding of both the efficacy of the intervention and its safety.

Natanson now takes a swipe at ARDS-NET (I can’t cope). He says that there is evidence in the literature that at the time it was performed, that clinicians world wide were titrating ventilation depending on severity of lung injury. So, some patients who normally would have had their ventilation volumes reduced as their lung compliance fell would be randomised to the high volume group, and those with good lungs were prescribed low volumes – this is completely contrary to “usual practice”. Supporting this theory, the expected mortality in ARDS at the time of the trial was similar to the intervention arm outcomes, and much less than the so-called control arm outcomes.

Again, this interferes with the interpretation of the results.


One of the biggest risks with this issue is the lack of a control for monitoring safety. In our high noise-to-signal environment with a high mortality, harms due to therapy can easily be misinterpreted as harm due to underlying illness. This theory may explain why expected mortality rates are higher in the “control” arm of trials such as the ARDS-NET trial and the Rivers’ EGDT trial.

So what do we do about this? Our trials generally target highly titrated therapies – drug infusions, fluid volumes, ventilation volumes, blood pressure targets just being a few.

Intelligent trial design is require for this – something to be discussed later in the program.

Todd Fraser from Australia wrote 05-31-2013 08:23:10 am
To round out a fantastic session, another massive name takes the stage – Luciano Gattinoni. I’ve been waiting years to hear him speak.

Gattinoni argues that there is great confusion in clinical practice – the disease process, the host response, the complications and effects of the therapies that we use. Perhaps we lose site of the primary issues, and we need to return to the idea of “supporting the patient, to buy time”.

He cites ARDS management as a great example of this. Nearly 30 years ago, Gattinoni himself was ventilating patients with a lung protective strategy, and scrubbing the blood clean of CO2. And yet during that time, complications from ARDS seemed to rise. How did we let this happen?

He raises an interesting point – how on earth did we progress medicine without the RCT? Have we forgotten our roots in basic sciences and common sense in the clamour for EBM?

He says this common sense approach has led to, for example, an understanding of why there is dependent collapse related to gravity in ARDS, despite the prevailing understanding at the time that ARDS is a diffuse pathogenic process. Additionally, the application of PEEP to counter this process came from observation and understanding.

Turning to PEEP, he says that he believes he used evidence of “recruitability” (based on CT scanning) at any particular PEEP level and found that the best was in the 10-15cmH2O zone. However, there is evidence that this differs for different patients, highlighting points raised by Charles Natanson.

He also reinforces Streat’s view that application of science is skewed dramatically by emotion and other non-science factors. Take Selective Digestive Decontamination, which has a supportive evidence base better than most interventions and is rarely used, with the rapid uptake of relatively unproven tight glycaemic control (and has now seemingly been shot down).

EBM, he concludes, has a role, but have we killed medical reasoning in the process? What’s wrong with Physiology based medicine? Epidemiology based medicine?

Todd Fraser from Australia wrote 05-31-2013 08:23:27 am
What a great way to set the scene for the rest of the meeting. Back after morning tea to dispel some of the myths in specific areas of practice.

Todd Fraser from Australia wrote 05-31-2013 08:24:50 am
Session 2

The Directors of Auckland City and Wellington hospitals now take control of session 2. Colin McArthur is no stranger to Australasian intensivists, while Shawn Sturland is yet another expatriated northerner finding a home in the antipodes. In this session, we’re told, we’ll be challenging the dogma of ICU research.

Todd Fraser from Australia wrote 05-31-2013 09:19:25 am
First up we welcome Charles Natanson back to the stage to follow up the issue of hot topic “Anaemia and transfusion in the ICU”

He points out that over 80 million units of blood are collected each year. This is a major industry. Most blood requires storage, with most countries having expiry times of around 7 weeks. He notes that if it was proven that age of blood has an impact on outcomes, the reverberations would be enormous.

Stored blood may impact on outcomes via a number of mechanisms, including iron use by bacteria (beneficial to the pathogen), breakdown lipids and other antigens that cause inflammation. A new cause is also proposed - cell-free haemoglobin may also scavange nitric oxide, with consequent complications such as vasoconstriction, both pulmonary and systemic, and intravascular coagulation)

A number of new trials looking at this issue are in process – ABLE, RECESS, INFORM, TRANSFUSE etc that will give some more information. APIRI has been finished, awaiting results.

What’s the evidence base? In terms of RCTs, incredibly small. The best evidence comes from large registries, though obviously these are limited by their retrospective nature. Nonetheless, the overall results (Wnag et al, Transfusion 2012) seems to favour fresher blood. The effect is consistently seen in major patient groups (trauma, cardiac surgery, paediatrics etc)

How big is the problem? It appears, based on this limited data, the number needed to treat is 97 to avoid one death. (95% CI 63-199). However, looking at less sick patients in this study, the NNT rises to nearly 10000! So, are we jumping at shadows?

He shows evidence to support the theory that cell-free haemoglobin might contribute to the “storage lesion” of older blood. Cell-free haemoglobin levels in dog models remain raised for several days after transfusion, and NO levels remain depressed over the same period. Similarly, there is evidence supporting the theory that the free iron is avidly taken up by the pathogens in a dog model of pneumonia.

Perhaps this can be measured using haptoglobin – this measure of haemolysis of the old transfused blood appears to correlate with these other effects. Perhaps in time this will become a marker of the impact of “storage lesion” in our patients.

Two more things I learned in this talk – there is a canine blood bank in the US (I presume donors get dog biscuits after they donate?), and dogs with cancer can have chemotherapy. Hmmmmm

Todd Fraser from Australia wrote 05-31-2013 09:44:52 am
Steve Webb up now to talk to us about the fundamental question in clinical practice – what does it take to change our practice?

He start’s by quoting Brad Power who says there are only 4 criteria that are positively associated with uptake by clinicians - Expensive, dangerous, useless and fun!

Choosing the right endpoints for a trial makes a massive difference. He says while he understands the criticisms levelled at randomised controlled trials, he feels compelled to defend its virtues. Parachutes, he says, are uncommon in the ICU.

Half of what we do is wrong, he says – we just don’t know which half. RCTs he says are the only way of unpicking this problem.

Why do trials?
• Buff the CV?
• Convert patients to airline tickets?
• Generate sex, power and money?

Perhaps not – to generate evidence to change practice and policy; to understand a mechanism about the disease is a bonus. For this reason we need endpoints.

He challenges the comments of those supporting intuitive practice, as he says they are effectively surrogate outcomes, not necessarily the outcome we want. Patient centred outcomes, he says, have far more meaning – survival, morbidity, experience – than surrogates commonly used in trials. Even then, a hard outcome such as 90 day mortality can be viewed as a surrogate for a long-healthy successful enjoyable life. We don’t always know this information from the outcome measures we use.

He demonstrates that a 90 day mortality endpoint is important as deaths attributable to critical illness continue out long after the patient leaves hospital.

He challenges the use of “survival time” because of the possibility that rather than reporting a benefit, the therapy is just converting patients into “a slow lingering death”, with no longer term benefit. This, he says, is an inappropriate endpoint for intensive care.

Surrogate outcomes have been plagued by disappointment – multiple studies have led to conclusions based on surrogates that have later proven detrimental when long term information was obtained. A great example is the DECRA study. In this study of decompressive craniectomy, there is substantial separation of the ICP measured in these patients, often linked with survival in traumatic brain injury. Other positive signs were the ventilator days, days in ICU etc. However, the patient centred end points all went in the opposite direction. Further evidence has been seen in the use of Growth Hormone, nitric oxide synthase inhibitors, hydrocortisone, dopamine and more.

So are surrogates EVER valid? He says we first need to prove patient centred outcomes are associated with that surrogate, something that is exceedingly rare at this point in time in ICU.

Composite endpoints try to get more bang for buck in clinical research. However, they too are plagued with problems, such as when two parts of the composite move in different directions, negating any benefit that might be shown. He says he feels having a primary and secondary outcome is a far better strategy.

A true phase 3 trial he says should contain patient centred endpoints.


So, what to do with EBM? It will always be a combination of intuitive management interlaced with EBM where possible. At present, there is too little of the latter.

He too alludes to the “Goldilocks Principle” of clinical medicine – not too much, not to little, just right. This is the classive U shaped curve of fiddling with physiological variables. The problem is working out what the shape of the curve is, and what factors influence the curve in individual patients

He closes by quoting Robert McNamara – there is a tendancy to make the measurable important, rather than the important measurable.

Todd Fraser from Australia wrote 05-31-2013 10:11:42 am
Finally, Professor Roger Lewis joins in to challenge the long held concepts in clinical trial design.

Current problems
• Phase 3 trials seem to fail more commonly than expected based on prior work
• A high rate of failed trials
• A high rate of failure late in the process, with consequent costs

So why are we having these problems?

Because of trial design, we tend to fix as many variables as we can, something not reflective of daily practice.

We also have an inefficiency in that trials that fail do not do so quickly, resulting in waste.

Here, he says, is where adaptive trial design holds some promise. He says this prevents problems where assumptions made at the start of the trial (which are fixed) prevent adaption of the study based on information that comes FROM the trial. This can make it more efficient, in terms of time, costs or trials that reach incorrect conclusions.

For example, if there appears to be benefit in one of multiple treatment arms, but not the others, randomisation can be adapted to concentrate on the target. This might reduce the number of patients required to be enrolled, or prevent the need for another trial based on subgroup analysis. This is very useful, for example, in dose finding studies – once there is a trend to a best dose, resources are committed to this area to get the best chance of a positive result. The same can be applied to identifying subgroups based on biomarkers.

This isn’t a free-for-all though – adaptions, criteria for doing so and the type of changes must be identified and tested a priori.

Problems :
• Frequent looking at the data
• They are very complex and require complicated statistical support
• They require significantly more preparation at the front end of a trial
• Potential diversion down an incorrect path based solely on chance
• Difficulty with “informed consent”
• Difficulty with ethics approval and funding (unclear goalposts)

He moves on to demonstrate how this might be effective in practice. His example dramatically increases the chances of developing a successful trial while reducing the number of patients required to undertake it.

Todd Fraser from Australia wrote 05-31-2013 10:23:35 am
Discussion for Session 2

• John Myburgh – Landmark trials seem to stop early, but ANZ trials seem to go to completion. Can we integrate a robust stopping rule into adaptive designs? Response (Roger Lewis) - Increasingly its felt that a trial that stops early is somehow flawed, but there is little statistical evidence to support this. Adaptive designs prespecify stopping rules to avoid confusion in this context
• Vineet - 90 day mortality misses the potential causes of mortality and the information that this could give us. Response (Steve Webb) – true, but at least death is a hard endpoint and difficult to mess up.
• Luciano Gattinoni – Don’t adaptive designs just “Play the winner?” Response (Roger Lewis) – the aim of the phase 2 trial is to establish if there is enough evidence to go on to a phase 3 trial. The answer should be yes or no, not undefined – this is a failure of the trial design.


And now its off to lunch

Todd Fraser from Australia wrote 05-31-2013 12:21:00 pm
Session 3 - Busting Sepsis Myths

What Bala Venkatesh doesn’t know about sepsis probably isn’t worth knowing, but its good he turned up anyway. In this session he's joined by local Dr Chris Poynter, the organiser of the recent Wellington Exam Preparation conference, who’s star is definitely on the rise.

Their mission? To keep three of the biggest names in ICU on track! Andrew Hilton, Jeff Lipman and the big daddy of Australasian research, Rinaldo Bellomo

Todd Fraser from Australia wrote 05-31-2013 12:56:17 pm
First up Andrew Hilton takes a well overdue swing at the prevailing practice of therapeutic drowning (otherwise known as fluid resuscitation) in sepsis.

Did he just use the words “goal directed” in an ICU presentation?

Andrew starts by giving a classic example of a horribly physiologically unstable patient. The diagnosis remains unclear, so what is causing the severe, constrictor dependent hypotension, and how do we deal with it?

He points to numerous international guidelines that recommend bolus fluids as initial therapy for undifferentiated shock, and for specific groups such as sepsis. Reviewing the literature, he finds only 4 with patient centred outcomes, and in only one was the use of fluid tested – the FEAST trial.

Hilton claims that fluids are treated like the parachute – so self evidently useful that an RCT is not necessary.

He explores the series of assumptions required for the use of fluids, starting at the end :

He challenges the widely held belief that higher cardiac outputs are equated with higher likelihood of survival, no matter what the cause of the shock. He refers to a number of studies that have failed to achieve this when cardiac output is artificially raised. In a similar sense, does improving cardiac output improve tissue perfusion? There is some degree of separation between the macrocirculation and the microcirculation, and a change in the former might not lead to a change in the latter.

Even if we do accept the premise that raising cardiac output will be good for you, do we really know that fluids will achieve this? And then if we believe that it will, are there other methods that achieve this more safely, effectively and cheaply (such as inotropes and constrictors)?

That doesn’t even begin to address what fluid and what amount, or how you titrate it (serum lactate, tissue oxygenation, macrovascular measures).

Clearly a number of assumptions are required, and we currently don’t know enough to be confident of our management.

He neatly uses echo (he had to get it in somewhere!) to demonstrate that hypotension in sepsis is not always due to overt hypovolaemia, nor does it always create a response in a clinical sense in septic shock. Myocardial depression is common in sepsis, more so than sometimes realised, and in this context fluid therapy may indeed be harmful!

Are fluid boluses harmful? Possibly, he says, moving to the FEAST trial. This landmark study found that severely septic children in Africa died more often when bolus fluids were given compared with a more cautious approach. While limited in some ways, there is still important information to be gained here – we need to question our reliance on bolus therapy carefully.

Todd Fraser from Australia wrote 05-31-2013 12:59:40 pm
Stay up to date with the latest by following #CICMASM13 on Twitter - Nickson, Flower, ICN and more

Todd Fraser from Australia wrote 05-31-2013 01:15:10 pm
You can see he’s been champing at the bit – Royal Brisbane’s Jeff Lipman is up next, on his area of interest, antibiotic management in sepsis.

First up, Jeff addresses the issues of definition in sepsis in ICU. He starts with VAP and indicates that the definition relies on multiple subjective criteria – so we’re all treating different things. VAP is quickly being replaced by a new clinical definition of Infection related Ventialor Associated Complications (IVAC).

Similar problems exist with Line related Sepsis, Ventriculitis and infective complications in burns, trauma and pancreatitis.

Picking these problems would be a lot easier if we had a decent biomarker of infection, but sadly none exist. Procalcitonin seemed to be the way forward but recent meta-analyses (eg Heyland et al) appear to suggest it will fail to live up to the hype – certainly you can’t use it to make decisions on starting antibiotics or not. We desperately need other markers coming through.

In terms of duration, he says the first decision is source control. Duration must be at least as long as it takes to get control of the source, so if there is undrainable pus somewhere, don’t even think about it.

Antibiotic duration is a tradeoff between under-treating, and generation of resistance. Some studies have demonstrated that if you get a second infection after prolonged courses, a repead infection in the same hospital admission leads to worse outcomes, suggesting resistant / selection of other organisms was occurring in individuals rapidly.

Does dual therapy help? Perhaps not! Maybe it just speeds up and broadens the development of resistance.

He suggests a week of treatment in VAP is usually enough. For CAP, 5 days, with at least 3 days fever-free. Severe UTIs need 5-7 days, and severe intra-abdominal infections (post source control) can get away with 4-7 days. Pseudomonas infections are slightly different and need about 10 days.

Staph aureus though is a sticky organism that requires a prolonged course due to recurrence. Most authorities suggest a 2 week course for invasive infections.

At this stage, there are no hard answers yet – more research is required.

Todd Fraser from Australia wrote 05-31-2013 01:39:54 pm
You know you’ve made it as a performer when you are known world wide by first name only – think Madonna, think Angelina, think Kermit – and then there’s Rinaldo…

Rinaldo’s brief? What are the blood pressure targets in sepsis?

He begins by redressing Steve Webb’s U shaped “Sweet Spot” described in a review in Critical Care and Resuscitation. The problem of course is identifying what threshold is important for each patient.

An additional issue is that in pratice we use specific cutoffs, which are always arbitrary, to trigger interventions – for example, “keep the MAP >65mmHg”.

I guess the place to start is baseline blood pressure. But what is that? Blood pressure in the doctor’s office? Or daytime? Or nighttime? There is a 15mmHg difference between the latter two – so where are the goal posts? The answer is, of course, unknown.

Some studies have been done where baseline blood pressure data is strong. In these studies, some early information suggests that the further the patient’s blood pressure goes from baseline (known as the “mean perfusion pressure deficit”), and for a longer duration, the worse the patient’s outcomes. The studies also indicate that most patients spend prolonged periods below their baseline. Is there a cause and effect mechanism here? Or just association?

The next step in the logic chain is that if we choose a higher figure, will this improve outcomes? And if we use a drug to get there, will this cause more harm than good?

Adrenergic agents, most commonly used for this purpose, are associated with some harm in sepsis – for example, pre-treating septic animals with beta blockers has a protective effect! In clinical practice, prior-use of betablockers also pretects among hospitalised sepsis patients. This appears to be a consistent effect across subgroups. Nonetheless, they said that about statins too until they did the proper trial.

Still, I’m not sure anyone will be brave enough to give beta blockers to a septic, unstable patient.

He presents data from his group of a sheep model of sepsis. Cardiac output increases quite markedly in these animals. If betablockers are administered 6-8 hours after onset of clinical sepsis, cardiac output still rises, but not by much, nor does MAP. Outcomes are not available yet, but this seems to suggest that we should be less afraid of the consequences of beta blockade in these circumstances. Is this just another fad? Further evidence awaits.

So in summary, we have no idea what the target, or best method of managing, or if one patient needs different settings to another. But before we get too carried away, perhaps we’ll be blocking it anyway!

Todd Fraser from Australia wrote 05-31-2013 01:45:20 pm
Session 3 Questions from the floor

• Bob Wright – In anaphylaxis, similar to septic shock in that it causes vasodilation and capillary leak, we use catecholamines to correct this. Is there a role for this? Response (Andrew Hilton) – not sure, there are obviously clearly volume deplete patients, they need to be made euvolaemic. However, yes, perhaps, and it highlights that once euvolaemic, other methods are important to consider rather than fluids.
• Vineet – does the quantum of bugs make a difference? Response (Jeff Lipman) – hard to answer, no research to guide.
• Bala Venkatesh – what are your thoughts on the studies that show benefit for Procalcitonin to guide therapy length? Response (Jeff Lipman) – yes, you probably can, but only if the control arm has a long course – if you use short courses already, procalcitonin isn’t going to help too much. No RCTs exist (except Chastre study) for limiting duration of antibiotics
• John Myburgh – What trial would you do to tease this out? Response (Rinaldo Bellomo) – would study patients on vasopressors, without cardiogenic element, targeting MAP 20% below baseline, 10% below baseline and baseline. He’d do an adaptive trial! Would also love to use +/- beta blockers in all groups, and use the adaptive trial to kill off the ones that don’t work. We still know so little, its amazing given this is a ubiquitous therapy

Todd Fraser from Australia wrote 05-31-2013 02:46:38 pm
Session 4

In the final session of the day, we welcome back local yokel Paul Young, and with him is Tony Williams. They are here to oversee the smashing of some more half truths and mythology in toxicology and renal failure

Todd Fraser from Australia wrote 05-31-2013 02:46:56 pm
Young gun from #FOAMed and Life in the Fast Lane fame Chris Nickson is up first.

Dogma 1 - Ever the dynamic speaker, Chris begins with Glucagon (or is it glucagone?) in calcium blocker overdose. He cites a review of the literature which shows almost no evidence to support its use. Essentially unblended, uncontrolled animal based, glucagon seems to increase heart rate but that’s about it. It appears insulin has a better effect profile, but largely the evidence for it is as tepid as that for glucagon. Just replacing one dogma for another?
Interestingly, older glucagon contained some insulin – was this the reason it was working in the first place?

Dogma 2 – can you give calcium for hyperkalaemia in digoxin toxicity? This has been traditionally a no-no because in theory digoxin toxicity causes high intracellular calcium, and may be worsened by calcium injection. This might lead to the contracted, immobile myocardium – the so called stone heart.

But is there evidence for this? Not really – in a study of digoxin toxicity, those who were given calcium (either intentionally or inadvertently) did not do as badly as would be expected.

Dogma 3 – paracetamol toxicity and the timing of liver transplantation. Currently, the need for liver transplantation has been guided by the King’s criteria. Nickson presents some data by Buckley et al demonstrating that while patients who are transplanted after meeting the criteria do better in the first 20 years, those that were not do better beyond the 20 year mark – a finding probably explained by the complications of immunosuppression and transplant.

So toxicology practice it seems is littered with poor quality animal studies, rumour, habit and circumstance – much like the rest of critical care really!

Todd Fraser from Australia wrote 05-31-2013 03:18:46 pm
Here he is again, the best thing ever to come out of Heidelberg (except perhaps Burgandy st), Rinaldo Bellomo busting up renal B.S.

So what does oliguria mean nowadays?

We’re better at sitting on our hands than we were, but there is still a brainstem response to fill the patient in most circumstances when the UO is persistently low.

The evidence, he says, suggests that most patients who are given fluid for low UO are fluid deplete, nor does giving fluid actually assist this. When it does work, the response is short lived.

So why is this dogma held on to so tightly?

The basic paradigms are held as true are :
• AKI is due to poor flow
• Fluid loading improves flow

Addressing the first of these, he demonstrates that in a model of extreme renal ischaemia (a renal artery clamp) led to almost indetectable changes in renal function. Even models of sepsis suggest that renal perfusion actually increases, despite rising creatinine and oliguria. He says the kidney is actually quite resistant to ischaemia – one of the great dogmas of critical care.

On the second of these, there is evidence that volume infusion in these circumstances makes little difference to the renal blood flow.

Consequently, he argues, further fluid boluses are highly unlikely to make any difference in patients with sepsis and oliguria.

It is noteworthy that when we give fluids the Cr goes down. This has been demonstrated frequently, and is often held as evidence that fluids improve renal function. However, this has been examined and appears to be completely explained by dilution effect.

Surely keeping the kidneys “wet” will at least be advantageous, or more importantly, a “dry” strategy will cause more renal failure. Wrong again. Rinaldo brings out the FACCT trial, where the dry group actually had better renal outcomes in terms of need for RRT. It seems water-logged kidneys don’t work very well, swollen and trapped within their capsules.

Excessive fluid balance has been linked with complications and morbidity in many patient groups such as elective abdominal surgery and ARDS, and in fact now is predictive of increased mortality as well.

Perhaps we’ve got this around the wrong way – perhaps as Mervyn Singer says, the kidneys are shutting down to protect themselves.

Maybe we’re already starting to change our practice – in SAFE, the first couple of days saw average of 2.5L of fluid given. In the CHEST trial, done 10 years later, in an almost indistinguishable cohort, the average was 500-600ml.

He concludes with the observation, “if AKI was fixed by fluids, AKI wouldn’t exist, would it?”. Enough said.

Todd Fraser from Australia wrote 05-31-2013 03:21:36 pm
Questions from the floor

• Manoj Sexana – how do you deal with the low-frequency event? Response (Chris Nickson) – this must be the role of collectives like CICM / ANZICS, to increase the net to catch events like this. Basically, every patient should be in a study or registry of some kind!
• Paul Young – is there a role for liver transplant any more? Bellomo jumps in – its virtually unheard of now in Victoria, Kings criteria are outdated.
• Bob Wright – normal saline induces a reduction in GFR due to hyperchloraemia. Does this affect the studies? Response (Rinaldo Bellomo) – most studies used Normal Saline. A recent crossover study demonstrated high chloride containing solutions reduce renal blood flow, while the “balanced solutions” do not – so possibly so. This will be the subject of a new trial to start soon run by Paul Young
• Geoff Dobb - should we be removing low urine output from MET criteria and CEWT scores? Response (Rinaldo) – tempting yes, but evidence can’t support that now, because possibly oliguria is still an important marker of something not being right. Easy study, should be done.


Yibbedah yibbedah, that’s all folks. We’re off for drinkies – its called the “Blind Wine Tasting session for a very good reason!

See you tomorrow.

Todd Fraser from Australia wrote 06-01-2013 07:28:08 am
DAY 2

What idiot thought that a fun run would be a good idea?

Well done to all those who were out at Sparrow’s Fart this morning.


Session 1

Steve Webb is back in this morning, and with him is another local Ben Barrie. Were about to find out we’ve been mucking up respiratory failure all these years.

Todd Fraser from Australia wrote 06-01-2013 07:30:39 am
Going back to where it all started, Luciano Gattinoni revisits a forgotten concept, extracorporeal CO2 removal. Prof Gattinoni introduced this process nearly 30 years ago

In the early days of ARDS (remembering it’s a disease of ICU, and that specialty is only 40 years old), patients frequently suffered what we now know as barotraumas. Recognising early on that this was due to high vent pressures, extracorporeal removal of CO2 was tried (way back in 1972!) with some success. However, early RCTs did not show any difference, with huge mortality in both arms.

He demonstrates that to remove adequate CO2, the technical elements required are much simpler – you only need lower blood flows and simpler circuits. Oxygenation is much more technically difficult.

After refining this technique, Gattinoni’s group was able to remove CO2 effectively in ARDS patients, aand that allowed them to lower pressures and volumes significantly.

He notes that when we measure compliance in patients with ARDS, we are usually measuring the bit that’s open, and the compliance of this is not that different to normal. This, he explains, is how the concept of baby-lung came into being. He says we need to forget about the collapsed unrecruitable bit of the lung and ventilate appropriately for the size of the remaining “baby lung”.

While it was not widely adopted, a number of units were regularly practicing ECMO for severe ARDS with better results than would be expected.

So where are we now?

He discusses the CESAR trial, where patients were either randomised to stay in their own hospital and receive normal care, or be shifted to an ECMO centre. He demonstrates some limitations with the trial, but it appears to show evidence of benefit.

Other evidence comes from the H1N1 epidemic, which showed markedly good survival rates compared with expectation. The sheer volume of cases also seems to have given us an injection of experience with ECMO, leading to even better results.

Controversially, he shows some data supporting it even in the context of bridging to lung transplantation.

Todd Fraser from Australia wrote 06-01-2013 07:53:40 am
Here’s the Prez again, champion of the regional ICU, Ross Freebairn. The ICU night nurses I know are going to love this…

Like many things, says Ross, paralysis seemed to drift into ICU from anaesthesia without any rigorous assessment of its virtues. In a 2003 survey, it appeared that while nearly 90% of patients received relaxants, this was mostly used in the first 6 hours of admission.

The benefits have been well known – particularly an improved ability to ventilate patients within safe pressure limits, and improved oxygenation by reducing oxygen consumption, particularly by respiratory muscles.

However, problems have been identified too, and there are a number of purported advantages in maintaining spontaneous efforts to breathe and cough. Other risks include inability to monitor neurological function, and the risks of circuit disconnection and other harms (corneal abrasions, PTSD etc).

The big worry though has always been the link to the development of neuromuscular weakness. However, a recent RCT in ARDS seems to have dispelled this concern.

He presents data that suggests that reduction of oxygen consumption is not as profound as we think, but P/F ratio certainly appears to rise when paralysis occurs.

Why would this occur? It remains a little unclear, but perhaps prevents breath stacking, reduced barotraumas / volutrauma or other as yet unidentified mechanisms.

He also addresses an issue confusing to many, whether it is the applied pressure or the total transalveolar pressure that injures lung. The argument has been that we often ignore the negative pleural pressure in spontaneously breathing patients that increases the transalveolar pressure. He demonstrates that when spontaneously breathing, limiting the tidal volume still reduces lung damage.

Todd Fraser from Australia wrote 06-01-2013 08:22:29 am
Back again, Prof Gattinoni is going to teach us how to optimise recruitment – manouvers or just more PEEP?

He starts with convention – the classic sigmoid shaped PV curve, with the lower inflection point marking the point of full recruitment, the upper inflection point marking alveolar stretch (ie increased pressure with no increase in volume), with the optimum ventilation range between them.

Is this real? Sadly no. He demonstrates that there are regions that are recruiting all along the inflation range – 10, 20, 30 or even 40cmH2O – even in normal lungs. The concept that lung is fully recruited at the lower inflection point is clearly wrong, he says.

Having said that, he also demonstrates that the higher the pressures go, the less value you get in terms of volume recruited.

Recruitment occurs during inspiration, derecruitment occurs in expiration, says Gattinoni. Why does it collapse? Lung weight is the most important. Regional hypoventilation and surfactant deficit may play roles, though he feels these are less important.

Lung does recruit through all stages of inspiration, but derecruits on the way down. CT slices from his studies show clearly that this can be prevented by using higher levels of PEEP. Quite impressive really.

He reminds us that the energy we use to inflate is partially used to overcome resistance to inflow, which may explain why collapse occurs at lower pressures than the pressures required to open them. In other words, higher pressures are required to open the lung than to keep them open.

By extension then, the lung that only opens at higher levels requires higher levels of PEEP to keep them open.

So really, the ideal PEEP becomes a trade off of best recruitability versus the potential consequences of high lung pressures.

Practically speaking then, how do we select the best PEEP?


He compared various methods of setting PEEP :
• CT guided
• Oesophageal pressure
• Stress Index method
• ExPress method
• LOVS study method
• ARDS-NET

None yet stands out. CT scan for him is available and appears valuable, but recognises this is not practical in most places.

He says a common sense approach is that the more likely there is recruitability, the higher the PEEP should be. How to tell this? This is the difficult bit. More work required in this area.

Todd Fraser from Australia wrote 06-01-2013 08:29:00 am
Questions from the floor

• Ben Barrie – how do you interpret the recent PROSEVA trial? Game Changer? – Response (Gattinoni) the mortality benefit was very high, this is unlikely to be replicated in real world practice. The patients might be a bit different too with a lot of viral pneumonitis patients. Howeer, he believes in severe ARDS it remains an important tool.
• Craig French – what is the impact of chest wall compliance on recruitability and does this change your strategy? Response (Gattinoni) – in the right position, extrapulmonary compliance can be decreased. Measuring the change in compliance measured on the ventilator is how you measure the recruitability, and this is largely the same in all patient groups. He highlights the predeliction of H1N1 for obese and pregnant patients – perhaps the decrease in extrapulmonary compliance influenced this, and using a blanket rule of plateau pressures of 30cmH2O may have resulted in under recruitment

Todd Fraser from Australia wrote 06-01-2013 09:08:52 am
Day 2 session 2 ICU quality – can we do it better?

That’s the question that George Downward and Bob Ure are here to find the answers to as they chair the next session.

Todd Fraser from Australia wrote 06-01-2013 09:27:36 am
We start with Darwin’s very own force of nature, Diane Stephens

Di, as many will know, is the first FCICM to practice in the Northern Capital, and went there as a brand spanking new graduate, something few are brave enough to do. She works in one of the most spectacular backdrops you can imagine, but it is the clinical landscape that can sometimes be more breathtaking.

Her initial mission is to discuss the gender imbalance in our specialty in Australasia. Importantly, only 18% of FCICMs are women, but as time goes on this may be addressed, as females represent over 30% of trainees.

Worklife balance though is no longer a predominantly female issue – more men are seeking a sustainable W-L balance.

Turning her attention to ICM training, she notes that we know that Intensivists, on site, in closed units, leads to best outcomes. Di should know this well – she’s one of the few people who can directly measure her own impact – she began at Darwin during the middle of a trial. This trial needed to be adjusted for her presence, as there was a 36% reduction in mortality following her arrival! She’s the only person I know who has their own Kaplan-Meier curve…

But are we doing enough? She raises an interesting concept – we use Intensivists in bolus doses – and sometimes there is both a delay to first dose and a prolonged interval between doses. Thus far there are studies that have failed to demonstrate a benefit of in-house night time staffing in the models of care currently used in Australasia.

Despite this, there are still concerns that higher intensity consultant presence may improve outcomes further. Her solution? Extended hours / evening shifts. The advantages being better continuity, better supervision of registrars, improved planning leading into nights that results in smoother clinical courses and better teaching of relatively junior registrars (with attendant positive experiences).

Controversially (perhaps deliberately?) she argues that the days of dual training are over – the demands and expertise of maintaining currency in ICU medicine requires dedication to a single specialty, she says. Sticking to our own primary, our own structures and picking appropriate trainees will deliver us the best practitioners.

Todd Fraser from Australia wrote 06-01-2013 09:50:05 am
Bundles are it and a bit at the moment – or are they, asks Prof Derek Angus again.

Why is there such passionate antipathy against bundles, checklists and guidelines? Many issues arise, including
• They are dumbed down practice - “Cookbook medicine”
• They fail to reflect experienced and nuanced practice
• Restrict autonomy of practice
• They may be based on dubios evidence
• I prefer my own evidence
• I shouldn’t be judged by them – in places they are being used to monitor performance
• They don’t work
• They don’t give me money, sex or power

But while there is truth to some of these, are they genuine reasons not to use them?

So why do we need them?

There is a huge amount of evidence that demonstrates a failure of clinicians to implement best practice based on available evidence. What are the reasons for this? Most clinicians genuinely want to do the best for their patients, so why such variation from accepted guidelines? Understanding this will clearly be useful in developing tools to improve compliance.

What is a bundle?

The concept is that a group of interventions are implemented together have important benefits to a group of patients, for example, a bundle for preventing VAP. The concept is not new, with the first record in healthcare in 1979.

Advantages – improves recallection of important interventions, and is a good memory cue / publicity tool. Sadly though, it makes working out which elements are effective very difficult, and may apply interventions inappropriately to groups of patients.

What about checklists? These are frequently used in other industries with high-risk procedures.

Checklists have ensured nothing is forgotten, and has empowered followers to have a voice – flattening the authority gradient that has been identified as a recurrent theme in aviation misadventure.

But do they make a difference?

At least in a study environment, there seems to be good evidence to support that they both improve compliance, and can improve patience centred outcomes. Good examples include Pronovost’s paper on Catheter infection prevention, the Surviving Sepsis Campaign, Surgical timeouts and so on. However, examining their impact is very difficult and most studies have badly flawed methodology (unblended, unrandomised, biased by temporal trends, sometimes incentivised) that limit the inferences that can be made. The Hawthorne effect also cannot be ignored.

Fascinatingly, the reported outcomes are better than the sum of the parts, even though they are implemented less well. How can this be? Does this “pass the sniff test?” Perhaps the checklist or algorithm itself is the more important bit – the effect on teamwork / flattened hierarchy etc – rather than the individual elements themselves.

Bundles and checklists on their own are not enough, as they still require multiple enabling strategies.

Todd Fraser from Australia wrote 06-01-2013 10:17:01 am
Following Prof Angus’ lead, Chris Natanson looks at the role of bundles in 3 common ICU problems – CVC infection prevention, VAP prevention and sepsis management.

The background to bundles is a natural evolution from passive recognition of harm events, through active identification, to developing a process to actively prevent them.

Is there validity in tracking process as a surrogate of outcome in healthcare? Maybe not. Maybe there is a more tenuous link between the process of care and outcome than we think, because of the incredible number of variables in the system, too many for us to control completely.

Essentially, bundles are an effort to ensure that all critical elements for any given clinical issue are delivered.

According to Nolan and Berwick (JAMA 2006) they should have indisputable elements with explicit rational, logical links between elements (that are additive and not antagonistic), and evidence based. A great example is Pronovost’s CVC infection prevention process.

What’s important, says Natanson, is that the implementation process is equally important, as it impacts not just on the elements of the bundle but on the culture of the unit itself – examples being teamwork, improved education, feedback loops to staff where they are more aware of outcomes etc.

He shows a Forrest plot demonstrating the effect of sepsis bundles resulting in a favourable outcome benefit on mortality. Is it important to tease out what the effective element is? In this study, they tried to do so and demonstrated that earlier, more effective antibiotic therapy appears to be the most important element, while others don’t have a clear benefit. This theory is certainly supported by the available data (Kumar CCM 2006).

He raises the valid concern about the reproducibility of studies becoming a thing of the past. This reinforces the point raised in the last talk about the criteria for elements of a bundle. In many bundles, he says, the elements do not meet these benchmarks.

But as Angus said in the last talk, does that matter?

Bundles do have their limitations. They can lead to ineffective elements being packaged together with truly indispensible ones.


Bundles also completely ignore the variability of individual patients.

So, he argues, we need to develop standards for the development of good bundles to maximise the benefits we obtain from them, and limiting inappropriate application.

Todd Fraser from Australia wrote 06-01-2013 10:34:54 am
Questions from the floor

• Alex Brown – why should we exclude other college primary exams? Aren’t these experiences necessary? Doesn’t it affect recruitment in regional areas? – Response (Di Stephens) : we should be aspiring to ensure the best possible intensive care in all departments, and her concern is that dual fellowships erodes time to stay adequately up to date and skilled.
• How do we resist the influence of maladaptive forces in development of bundles – eg industry, beauracracy etc. Response (Charles Natanson) – we need vigilance, but we also need to remember what the guidelines are intended to be – advice rather than enforceable rules. (Angus) – guidelines and bundles are not the same, as Natanson said. He acknowledges that past bundles and guidelines had their problems, but this process was getting better. Interestingly, this process was now driving research – identification of areas of weakness was a strong driver for new research targets
• Geoff Shaw – uptake of technology is lowest in ICU compared with other specialties, does this mean we are not engaging with industry enough? Response (Angus) – Correct, ICU is very slow to adopt technology. Multiple reasons though for slow adoption, including funding and regulation.

And with that, its off to lunch!

Todd Fraser from Australia wrote 06-01-2013 12:32:52 pm
Session 4 - Process of care in ICU

If there is anything intensivists pride themselves on its process. I hope Frank van Haren and Dick Dinsdale are going to stand up for us.

Todd Fraser from Australia wrote 06-01-2013 12:59:32 pm
Derek Angus is back, talking about regionalisation, outreach and telemedicine.

In short, ICU is an expensive and limited resource. Access to ICU is currently variable, and there is a likely cap on the availability of these resources. We might not be able to spread the current level of care to all people. New solutions to overcome these problems are needed.

Telemedicine and outreach ICU are two potential solutions to this problem.

There are about 6 million adult admissions to ICU in the US annually. The mortality is 13% and costs about 1% of GDP.

Clearly, having enough Intensivists in all areas is not just dominated by numbers, but by other factors such as lifestyle and location. Attracting intensivists to some areas will always be difficult – again, solutions to overcome this are necessary.

So if we can’t have that level of care everywhere, can we reorganise our service delivery so that each patient has access to the right level of care? This system has been demonstrated effective in its application to trauma centres.

Unfortunately we have limited ability to control some of these factors – trauma occurs everywhere; surgery gone bad occurs everywhere.

Do higher volumes of patients make a difference to performance? Evidence from Kahn NEJM 2006 certainly suggests that mortality falls with higher throughput units.

How big is the problem then? How many patients are treated in regional hospitals? In the US, about 40% of all ICU patients are admitted to the small units, a figure similar to that seen in Australia (ANZICS CORE data).

Perhaps pre-hospital decision rules can simplify this and redirect patient traffic to appropriate hospitals – early evidence in Seattle that this might be effective.

What about telemedicine? Telemedicine has potential advantages. Studies have been performed to explore the effect, with mixed results. Lilly (JAMA 2011) found some impact. Clearly there is such variability in operating rules, systems, patient cohorts and so on, so that interpreting these results is incredibly difficult.

Perhaps its difficult to show a difference because the majority of our impact is not a major intervention (eg picking the diagnosis using telemedicine) but a lot of little interventions (such as enforcement of “housekeeping”) that makes a difference to patient outcomes.

Finally, evolving from the SARS epidemic in Canada, a program was introduced drawing all ICUs within a unit together to share processes, quality of care initiatives, etc. Perhaps this level of collaboration in itself can improve outcomes.

In conclusion, ICU availability will always fall short of need. New multipronged approaches are needed that deliver the right care to the right patient at the right time. Telemedicine, regionalisation and outreach programs may be part of this.

Todd Fraser from Australia wrote 06-01-2013 01:31:54 pm
Dr David Galler is up next – to have real impact, we need to fix the system we work within.

“We live forwards, but we understand backwards”. With that he begins looking and how we can overhaul the way we practice ICU to take the specialty forward.

Major revolutionary changes in practice are not uncommon (think antibiotics, think, germ theory, think chemotherapy), and we should be proud of what we’ve achieved, he says. But we still have some major work to do.

Benefits of our efforts follow the laws of diminishing return, but harms appear to increase linearly – finding the point of best return can be difficult.

So how do we review the systems to get the most out of them?

Driver diagrams – identifying primary goals for an organisation and the mechanisms to get there, are important, and can efficiently direct the efforts and resources.

Sharing responsibility for outcomes with the stakeholders involved in the patients’ care provides motivation for performance.

While information from clinical trials has been valuable, we will continue to be dogged by a lack of clear helpful evidence for most of what we do. How do we deal with this?

We have the opportunity to design the systems we want that achieve the outcomes we desire.

A final thought – perhaps it is helpful to imagine we are practicing 100 years ago, and that we are looking back at the standards of care we practice now. What will we think of our practice, the misunderstandings we currently hold? Perhaps this view will help keep us humble in our approach to current practice.

Todd Fraser from Australia wrote 06-01-2013 02:34:45 pm
Back from afternoon coffee.

Helping find some untruths in the field of neurointensive care, here’s Brisbane’s Rob Boots with Wellington local, Alex Psirides.

Todd Fraser from Australia wrote 06-01-2013 02:34:57 pm
First up is Oli Flower talking about iconoclastic views of neurointensive care

What to do when entering an evidence free zone? Flower explores some ways of approaching this using a case example, illustrating how difficult it might be in predicting outcome. IMPACT TBI is an example of a calculator that aims to make this prediction easier. CRASH study group have another, and the results are often fairly well aligned.

Where do you sit on the goldilocks principle? He describes a familiar clinical situation of anaemia, with theorhetical reasons both for and against transfusion, where the evidence isn’t completely helpful at guiding a decision.

Next on the hit list is ICP monitoring – he discusses the BEST TRIP study that bravely took on one of the biggest dogma’s of neurocritical care. While a very provocative finding, Flower argues there is enough differences in context to disable validity. So where does that leave us? History, theory and dogma, it seems.

Interestingly, DECRA had profound effects on ICP, but no effect on patient centred outcomes.

Todd Fraser from Australia wrote 06-01-2013 03:03:20 pm
Ex-prez and crowd favorite John Myburgh is here carrying a big club, and its got decompressive craniotomy written all over it.

Self deprecating as always, he congratulates the audience for sticking around. Awwww….

He begins by setting the scene – the exclusion criteria for an ICU bed in South Africa when JM was an intern were considerably more stringent that they are in Australia today. Anyone with traumatic coma was automatically excluded.

In the time since, raised ICP has generally been accepted as the major pathological entity of TBI.

He describes the Three Tier (BTF guidelines) method of brain specific management in ICU, and indicates decompressive craniectomy sits up high on tier 3.

He contends that the major factor that has impacted on survival is pre-hospital care. What happens after possibly has less impact than we’d like to think.

Myburgh returns to basic principles in the Monro Kelly Doctrine, suggesting its not a pressure problem but a volume problem. The swollen brain has no where to go. So the concept of decompressive craniectomy certainly makes sense.

The trick is picking the right patient. Indications, timing, trigger and technique are all currently debated. In addition, what endpoint do we consider acceptable, as mortality might not be the best measure of success.

So down to tin tacks – does it work?

Reviewing the evidence, he cites a chinese paper that showed bifrontal decompression had bigger impact on ICP than limited flap. This is the basis of the decision on method in DECRA, a jewel in the ANZICS-CTG crown.

DECRA itself, published in NEJM in 2011, randomised 155 patients under 60, less then 72 hours from injury, with a CT trigger of diffuse disease, to bifrontal flap or standard care. The impact on ICP was profound, but the patient centred outcomes were horrible – patients shifting from dead to horrifically brain dead survivors.

So why doesn’t it work? Perhaps reexpansion creates shear forces resulting in awful injury. Perhaps the damage had already been done, and DECRA just prevents them becoming brain dead. Either way, its not a good outcome.

The RESCUE trial is coming, looking at its role as a rescue therapy – patients who have ICPs in the face of maximal therapy – time will tell.

An RCT done in 2013 (Honeybul) has failed to show any benefit for its use.

As an aside, is there any role in non-TBI causes of raised ICP? Not really – triumphant case reports aside, there is nothing to support its use. In MCA infarction in young people, three small trials exist (DECIMAL, DENSITY, HAMLET), but a summary of all is that it doesn’t work. It should be noted that the numbers are very small.

In conclusion, he muses :
• Outcome is more related to factors we cannot control, than factors we can
• Genetics may play a part
• ICP is a marker of severity, and dramatically controlling it makes no difference (in this case)
• Treating ICP comes at a cost – and that cost is born by the patient and their family
• Do what we should and not what we can
• Beware of being a numerologist

Todd Fraser from Australia wrote 06-01-2013 03:18:20 pm
Questions from the floor

• Bob Wright – isn’t CRASH a critically flawed paper? Response (Oliver Flower) – there were significant methodological flaws, and while the answer on steroids remains unclear, the burden of proof is on demonstrating it works before its implemented (nice answer!)
• Rob Boots – is it reasonable that we still to rescue decompressions? Response (Myburgh) – this question was not answered by DECRA which did it proactively. The new RESCUE trial will answer this, but JM suspects that refractory ICP signals time to give up (in context of course!)
• Bob Wright – is it possible that the dye is cast by the time Decra is done at 48 hours? Response (Myburgh) – possibly. What to do if your caught in that position? Need to make a judgement based on other factors
• Rinaldo Bellomo – two large RCTs on TBI now in progress examining if inflammation evolving causing cellular oedema might be contributing to poor outcomes. Perhaps the vault is offering protection by preventing extreme swelling – preventing cellular separation and so on. Response (Myburgh) – interesting concept that has been raised before, but hard to imagine.
• John Dyett – is there any validity to the multicompartment theory, particularly with decompressing the thorax to influence ICP. Response (Myburgh) – interesting thought but likely to result in decompression through the falx and the foramen magnum –bad news.
• John Evans – often see haemorrhage after Decra. Could it be that the neurones stretch more than their blood supply? Response (Myburgh) – yes, probably, and this is more likely with the limited resection for some reason. This is a perfectly plausible idea

And that’s that. Now its time for the dinner, where a bunch of new stars get the gong, including Crit-IQ contributor Rob Bevan. Well done Bevo!!!!!

Todd Fraser from Australia wrote 06-02-2013 07:28:07 am
DAY 3

Can anyone spell antioxidant? Congratulations to those who actually made it here this morning.


Session 1

It’ll be spewing forth from both ends in this session. Di Stephens is back (I need a panadol) and with her is Geoff Dobb from the other side of the continent to talk about the ins and outs of ICU.


Who else would you want to hear talk about feeding the critical care patient than Gord Doig?

Understanding a bit about regulation of food intake in normal subjects may give some insights into feeding in the critically ill, but we still have very limited understanding of this. Its clear though that hunger and nutrient intake are not linked – many people are still not satiated even after adequate calorie intake – hence the obesity epidemic.

Is the hunger response – and associated hormonal response – protective in some way?

Malnutrition is defined as wt loss > 2% in one week or 5% in 1 month, and insufficient energy intake for >5 days. The consequences of malnutrition affect every body organ system. Multiple studies relate malnutrition to poor hospital outcomes, with higher mortality, longer hospital stays, more infections, worse rehabilitation prospects and so on.

In critical illness, about 40-50% of our patients are malnourished in most studies, with higher rates of infection and complications, and even death rate. Even if they were well nourished, the majority lose up to 15% of lean body weight during a 3 week stay in ICU, even with adequate intake. Average loss of protein can be 1.2% per day.

In short, even well nourished patients are becoming malnourished in the ICU.

In a large observational study by Heyland et al of ventilated patients, malnourished patients particularly benefit from better nutrition – with reduced mortality and morbidity – but overall, all patients do better.

In bone marrow transplantation, providing TPN prior to and during transplantation results in prevention of malnutrition and this translates to meaningful patient centred benefits. But does this apply in critical care?

The available evidence is not strong. However, it does appear that when studies have shown a benefit, it is strongest when feeding is started early.

In summary, we still have much to learn. Malnutrition is definitely not good for you, appears associated with harm, and appears to be preventable with early nutrition in some patient groups.

Todd Fraser from Australia wrote 06-02-2013 07:49:26 am
World reknowned scatologist Rinaldo Bellomo is back, but after a big night out, it looks like he’s going through the motions.

GI dysfunction is reportedly very common in ICU, with diarrhoea common, along with high gastric aspirates second, and constipation occurring in up to 15%

We know little about this issue in ICU, its causes or its complications, let alone appropriate therapy. A pilot study has been done to look at this at the Austin.

The Bristol stool chart remains the most commonly accepted qualitative measure of stools. Overall, formed stools are a rarity in ICU, and iatrogenic diarrhoea is the most common problem. Interestingly, post-ICU diarrhoea is also fairly common in these patients. Despite this, the most patients do not pass a stool on any given day. This means the most common state is “constipation” but the most frequent problem is iatrogenic diarrhoea

Is “constipation” (failure to defecate for >3 days) a problem? Or are we mis-defining non-defaecation as constipation. In many of these patients, they don’t actually have a full rectum and a “desire to go”, so is it a big deal? It really only becomes important when it causes mechanical problems (eg Ogilvie’s syndrome, discomfort). Given that the most common abnormality is iatrogenic diarrhoea, perhaps we are overdoing our preventative measures.

Few studies shed light on this. We certainly have effective ways of making people poo, but whether or not this actually benefits patients remains unclear.

Better research is required.

Todd Fraser from Australia wrote 06-02-2013 08:04:53 am
Questions from the floor, session 1

• Geoff Dobb – why do sick patients get hungry, is this an adaptive response? (Response – Doig) – it has been proposed that we don’t eat in illness for a good evolutionary reason. However, nothing in the literature suggests that this is a beneficial effect. Not everything is of evolutionary benefit.
• Bellomo – what about EDEN study – (response Doig) feeding started early in that trial, where low and high doses of feeding were used. He says the trial should at least convince you to feed a bit. At this point it doesn't suggest you should feed more. Its like dialysis, we don’t know the dose, but we think it is of benefit, but its probably more effective early!
• Matt Bailey – what about protein balance (Response - Doig) only one trial. Feed formulae is fixed and linked to energy, which limits the inferences available from observational studies. Recently finished a trial on exactly this, results coming.
• Geoff Shaw – nutrition is effectively a bundle of care! Do we need to treat it as such, do we need to look more carefully at the components. (Response - Doig) – as mentioned, the study is coming (at least on protein)
• Di Stephens – is there any evidence for anti-narcotic medications in critically ill for avoiding constipation? (Response – Bellomo) No studies at present – great project for a trainee! As mentioned though, we don’t know if non-defecation is actually a problem! He believes we pay far too much attention to it, rather than asking the important question.
• Breast fed babies have highly variable bowel patterns. Some of them poo incredibly infrequently, and paediatricians seem to treat them expectantly, and most do very well (Response - Bellomo) – this highlights how little we know about what is normal, desirable and pathological in ICU patients.
• Matt Bailey – what feeds contribute to GI dysfunction. (Response - Bellomo) – in the information that they have, there doesn’t appear to be a link with different formulae. Administration of laxatives probably greatly out-weight any other effects of the feed. Its not surprising that we don’t poo when there is almost no roughage in the feed.
• Geoff Dobb – what about fibre? Why is it there? There is almost no evidence. Should we be using it? (Response – Bellomo) – points out that Dobb is actually published in this area! Agrees, little evidence. Perhaps would reduce diarrhoea, but could also see a situation where, when exposed to narcotics, electrolyte imbalances etc, that this could create fibre bricks. Completely unknown.
• John Green – has certainly seen some complicatons of constipation during his career, so is more circumspect about it. Maybe laxative induced diarrhoea is worse when it unclogs a blocked up bowel. Should we start laxative earlier by preventing constipation in the first place, lessening the followup diarrhoea. (Response - Bellomo) Again, little is known, but appreciate the problem. Like most things needs to be individualised (eg spinal injury different to severe pneumonia different to brain injury) Again reinforces the need for further studies. A method for differentiating non-defecation from true constipation is something we need to figure out.

Todd Fraser from Australia wrote 06-02-2013 08:33:40 am
Day 3 Session 2 - Education in ICU

And at the final turn, incoming CICM exams chair Mary Pinder and legend John Myburgh are here to look at the meducation revolution.

Todd Fraser from Australia wrote 06-02-2013 08:56:59 am
First up is former exam chair Bala Venkatesh who asks the question – what is it that we should be teaching now?

Perhaps we need to revisit the teaching of clinical skills, he says. We’ve evolved to numerologists (a common theme at the meeting) and perhaps we’ve lost our way a bit.

He begins with a fascinating review of the evolution of clinical history and examination skills. Through the late 1800s we saw the development of laboratory and radiology, which improved some diagnostic capabilities, but began to remove the doctor from the patient, a trend that continues today.

Interestingly, there are now attempts to examine the value of the clinical signs themselves – a series on this has been regularly appearing in JAMA for years. This has possibly had the inadvertent effect of down-regulating the perceived importance of “patient-centred medicine”, and promoted “test-centred medicine”

Is there real evidence that clinical skills have deteriorated? Frequent patient complaints and patient misadventure investigations have illustrated the loss of these skills. A number of surveys on clinical skills appear to reinforce this.

Bala presents further information questioning the validity of test-based medicine, and highlights that history and examination faults allowed the errors to happen.

He raises some concerns about current trends :
• “Board rounds” not ward rounds
• Telemedicine – can you really manage patients by remote control
• ICU robots – just plain weird
• The MET team revolution – leads to obfuscation of responsibility by home teams
• Protocolisation – rather than considering carefully the patient’s issues, we respond without clinical reasoning

The college continues to place importance on this issue – a bad fail at the hot case level results in overall failure, regardless of performance on the rest of the exam. A review of performance at the exam over the years confirms that clinical skills continue to weaken.

In finale, he calls for a return to an emphasis on trainers to promote clinical examination skills, and for consultants to lead by example.

Todd Fraser from Australia wrote 06-02-2013 09:28:03 am
Up next is Chris Nickson, “social medialite” extraordinaire – is the internet the education way of the future? (If you’re reading this then the answer for you is probably “YES!”)

Chris begins by reflecting that we must be doing much good already because we turn out some extremely high quality doctors. However, he points out that we provide education extremely inefficiently, in that people are developing resources and our inability to share results in duplication, triplication…

Chris highlights some of the benefits of technology to share educational ideas.

He reviews some of the internet resources available to clinicians, including EmCrit, SMACC, ICN, LITFL and of course, Crit-IQ.

Spreading ideas and allowing collaborative design of new techniques is a major benefit of this. He cites and example of Delayed Sequence Intubation, a technique designed by Scott Weingart, but refined over time by the social media conversation that followed.

Online learning also allows better use of face to face time, improving the efficiency of education, something he refers to as “Flipping the classroom”

Pointedly he notes that we need to get involved and contribute to the discussion, or at least, be aware of it, because young doctors certainly will!

Are there downsides?
• Fear of technology
• Information overload – how do you find what you need
• Filter failure – unregulated information
• Copyright
• Patient confidentiality
• Horses for courses – is it right for all?
• Potential for manipulation

Google FOAM is a new customised search engine in google to assist looking for the right information.

Todd Fraser from Australia wrote 06-02-2013 09:57:31 am
And finally, one of my favorite presenters, Mater Brisbane’s John Morgan, presents a view from someone with more than 30 years of specialist practice. I’m sure he’s about to do his world famous Eeyore impersonation!

JM begins by discussing his concern at his own work life misalignment – shrinking superannuation and expanding prostate. Enough said.

He begins by taking us back to the origins of ICU in Australasia. We developed our own society, own exit exam and own endorsement in the specialty.

Resources were scarce. They worked hard, they had clear rules, and they did their best.

Times have changed – equipment gets thinner as the patients get fatter. Expenditure goes up exponentially, way out of proportion to population growth. The casemix has changed and there is a massive growth in HDU type work – where did this all come from?

The culture has changed too – he addresses the way Baby Boomers, Gen X, Gen Y and now Gen alpha approach ICU differently, from both a consumer and service delivery perspective. He cites the interpretation of futility as a classic example, with the focus on the rites of the individual completely overwhelming those of the collective.

He argues this effect is also seen in admission criteria – there is no such thing as an inappropriate patient; when we put barriers up, we’re seen as adversaries, and we end up just giving in.

Another example is the MET system, he describes as a “VMO sleep preserver” (something that got wide appreciation from the audience), resulting in ward deskilling and increasing job dissatisfaction in ICU.

Next he tees off at the insidious impact of private hospital work, a culture he sees creeping into public practice now too. The default call is to send anyone too difficult to the ICU, a process he describes as a “reopening” of the closed units (bed and breakfast patients).

He rounds on those who he feels are taking a “parachute” out of clinical practice, and bemoans the lack of example and leadership provided to our junior staff.

Fatigue is up next – and the loss of training opportunities that this creates.

He views us now as the hospital cleaning service.

Strangely enough, people continue to train in ICU, and there will be fewer jobs. The problem this creates is that this will result in dropping off of trainees (as there are no jobs for them), and service provision will increasingly shift back from senior registrars to senior staff. Will we have junior-seniors?

Valiantly, he looks for the bright side…

Todd Fraser from Australia wrote 06-02-2013 09:57:58 am
Questions from the floor, session 2

• Chris Poynter – we work within a different system now, so is it reasonable to fail people based clinical skills? (Response – Bala) There is still value in clinical signs, and people don’t fail for missing things that are hard to find, they are missing major problems like hemiplegia. The ability to detect problems and develop differential diagnoses is sadly lacking.
• Mary Pinter – what protections are there for patient confidentiality (Response – Nickson) – onus is on contributors to behave responsibly.
• Bala Venkatesh – doesn’t agree that peer review is adequate. A book would have far more oversight than social media. (Response – Nickson) – doesn’t claim FOAMed have the rigor as a book or a journal, but does believe that this is an extension of “corridor conversation”, something we all use to learn and share ideas. FOAMed and the internet are simply facilitating this.

Todd Fraser from Australia wrote 06-02-2013 09:59:15 am
Convenor Michael Reade concludes the conference. Next year’s CICM ASM is coming to you from Brisbane, where the topic will be "Risky Business - making the right decisions where there are no right answers"

Well that’s a wrap everyone. Its been a blast, particularly the Antarctic wind out there today. We hope you’ve enjoyed the blog coverage – please feel free to leave your comments below.

Paul Davies from Australia wrote 06-03-2013 09:10:03 am
Thanks Todd, fantastic wrap for those of us who weren't there.
Paul Davies, Melbourne

Sean McManus from Australia wrote 06-03-2013 09:38:00 pm
Thanks Todd, a great read and more motivation for those of us studying to pass the quiz and join this remarkable group of clinicians.

Terrible FOMO though, wish I could have been there.