Blog

Todd Fraser wrote 10-25-2012 03:36:12 pm
So, here we go, the first session of the 37th ANZICS / ACCCN Annual Scientific Meeting, the keynote speeches.

Warwick Anderson – Translation of research into practice
This session was a very interesting way to start the conference. ANZICS has been strongly involved with research under the banner of its Cliinical Trials Group. Prof Warwick Anderson, Chief Executive Officer of the National Health and Medical Research Council (NHMRC) outlined some of the difficulties translating research into measurable health outcomes. To address this, the NHMRC has now set up the Research Translation Faculty.
We practice in a world where healthcare and research is increasingly costly, and funding for research is falling. Interestingly, he suggested that initial clinical research is reasonably well funded (though only 20% of grant applications are successful), but commercial uptake on these findings is lacking, and further support for commercial development may be required.

Todd Fraser wrote 10-25-2012 08:02:27 pm
Gordon Rubenfeld – The global burden of critical illness
The second keynote speech for the session was given by Prof Gordon Rubenfeld, Chief of the program in Trauma, Emergency and Critical Care at Sunnybrook Health Services Centre in Toronto. Professor Rubenfeld spoke of the global burden of critical illness. He points out that we have only limited data on the epidemiology of critical illness world wide. Why? There are numerous reasons, but the vast difference between resources, pathology and definitions contributes to the dearth of knowledge. He uses Acute Lung Injury as an example – you can’t have ARDS without ventilator support, so even if the incidence of the cause is the same, the incidence of ARDS will be different between countries. Interesting thought. Similarly, in many countries, many patients will never reach an ICU, either because of accessibility or resource limits, making epidemiological research impossible.

Depsite these limitations, an attempt to quantify the burden has been made –
(Lancet 2010; 375:1339-46). The figures are staggering – a worldwide number of ventilation of over 13 million, 5.5 million cases of ARDS, 15 million cases of severe sepsis and over 5 million deaths. So the answer to his topic is “enormous”. Not surprisingly, the vast majority of this burden is born by low income countries.

Attempts have been made to prioritise interventions that improve overall health care, and Rubenfeld uses Ghana as an example. Very few of the interventions, not surprisingly, on a cost-benefit analysis, are related to critical care! He presents data clearly demonstrating the worsening gap between increasing healthcare costs globally, and the likely worsening of funding due to the aging population.

Some efforts to improve critical care in low-income regions have been made, a number of guidelines existing in the literature. Interestingly, many of the recommendations centre on improvements in the co-ordination of care. This is not that dissimilar to higher-income countries in the form of intervention bundles. Development of cost-effective alternatives to strategies that work in developed countries, training and collaboration / sharing of best practices are also important.

Pointedly, Rubenfeld also indentifies populations within “first world” countries such as Australia and Canada that reflect similar issues, such as indigenous, homeless and other communities.

Hearteningly, the intensive care research community is beginning to collaborate, with a fantastic example being the recent H1N1 influenza epidemic.

Rubenfeld challenges us by raising the question “who cares about critical illness if I don’t have clean water”. He touches on the politically volatile topic of rationing of healthcare. He challenges the widely held belief that first world clinicians don’t already engage in rationing, hinting that we need to bring this discussion into the public domain.

Nonetheless, he points to data that speculates that we are poor at identifying survivors from non-survivors, making “rationing” incredibly difficult. Additionally, most patients want to go through ICU, with most people still accepting a 10% chance of returning to their current level of function.

He finishes with his favourite quote - “Things that can’t go on forever…don’t.” Herbert Stein

Todd Fraser wrote 10-25-2012 08:05:21 pm
The ANZICS conference is a bit difficult for one person to cover because there are concurrent sessions, so I chose to go to the paediatrics fluid session, largely because Kathryn Maitland was presenting about FEAST. Neil Orford refers to her as a "Research Rockstar" - I think he's got a crush.

Todd Fraser wrote 10-25-2012 08:05:34 pm
Marino Festa – Fluid resuscitation in PICU

Marino starts by reviewing the classic Starling physiology theory, and then reviews a new data that suggests this is far too simplistic. He reviews the findings of the SAFE TRIPS stuffy that followed the famous SAFE trial – the most interesting finding being that fluid choice was mostly dictated by regional habit rather than any other factor.

Marino briefly reviews the complications of fluids, quoting a bounty of evidence that points to worsened outcomes associated with increased fluid balance, while acknowledging that there is an association, rather than a cause-effect.

He also alludes to an often ignored factor in tissue oedema, lymphatic flow. Interestingly, lymph flow is clearly inhibited by high CVPs.

Marino demonstrates that in sepsis, a disease characterized by microvascular obstruction, other neighbouring vessels may dilate and become leaky to compensate. This increased tissue fluid may be adaptive, and with our fluid therapy, we may initially augment this, but as time goes on and oedema increases, we may cause harm by increasing the diffusion distances between vessles and cells.

He then reviews Greg Martin’s interesting work in ARDS using frusemide and albumin – drying out the patient is associated with better pulmonary function.

So the problem is too much interstitial fluids. Marino suggests that its unlikely that the fluid type is the major factor, and goes on to suggest that other “magic bullets” may eventually be found. These might include regulation of the microvasculature flow (dobutamine, immunomodulation), direct compression and elevation, and perhaps CVP regulation.

Todd Fraser wrote 10-25-2012 08:14:03 pm
Kathryn Maitland – The Feast Study

Kathryn begins by presenting a video clinical example – nothing short of horrifying – and highlights the burden of this disease in resource-deplete areas. She reviews the background knowledge about severe malaria, and identifies that the known mortality was highest in those with acidosis, about 15%, while those with severe anaemia alone (defined as Hb <5) had <1% mortality.

As part of the lead up to FEAST, the investigators looked for evidence that highly acidotic septic children were hypovolaemic. They found that patients who received fluid boluses seemed to improve a number of variables such as acidosis, heart rate and blood pressure. This all appeared to point to the potential benefit of bolusing fluids to these very sick kids, hence the FEAST trial. They also looked at fluid type, and at the time there was limited data, which if anything suggested albumin might be a bit better.

She points out that the often recommended volume of >40ml/kg bolus was based on a study of about 40 children, only 9 of whom actually received that amount! This highlights the ambiguity of current recommendations, and the need for better evidence.

FEAST was conducted in East Africa – Kenya, Tanzania and Uganda. They strove to find if a bolus was better than not, and if so (which they expected), albumin was better than saline. She reinforces that the trial was designed to be “real world” – the lack of diagnostic classification in the results is largely attributable to this. Similarly, the trials were done in centres without access to western ICU level care.

Its worth noting there was a lot of training involved prior to the trial. It would be interesting to measure the impact of this factor alone.

The trial studied children with sepsis, reduced perfusion and one of respiratory distress or impaired consciousness.

Midway through the trial, they increased the fluid boluses in an attempt to ensure that the trial avoided futility, as there was a concern that 20ml/kg was too little.

Blood transfusion was given after an hour in the intervention arms, whereas the control arm received them earlier, something that varied slightly from the intervention arm. Does this explain the results?

The study was very well adhered to – almost no protocol violations.

Kathryn counters the concerns that this trial is really just a malaria trial, highlighting the significant proportion in whom the diagnosis was sepsis.

She shows a video of the health workers telling the camera what they think the results will be. Universally, they are convinced that bolus dosing improves outcomes. Of course, the trial demonstrated an INCREASE in mortality in the bolus group. This was an enormous surprise.

No matter what subgroups you look at, all appear to be worse off after fluids.

Kathryn is at pains to stress the application of the trial is context specific, but raises clearly concerns in other circumstances.

Questions from the floor
• Ian Seppelt – Quotes John Myburgh - "The strongest level of evidence is performed by a believer who finds the opposite to what they expected" - is there equipoise to do this trial in our context? KM stresses that there was a strong belief that boluses were going to help not hurt, so perhaps we should be considering it!
• Marino Festo – did you measure the haemodynamic response? KM - Yes – but you’ll have to wait until tomorrow!

Todd Fraser wrote 10-25-2012 08:17:29 pm
Rounding out the session, and the day, is Michael Yung, Paediatric Intensivist from Melbourne, who talks about fluid resuscitation in DKA for kids.


Michael Yung – DKA resuscitation

Michael makes an interesting initial observation that fluids alone reduce serum glucose by almost 50% in the first couple of hours.

Cerebral oedema remains the most important cause of harm and death in DKA in kids. Predictably, the highest risk is in the sickest kids. Mortality is also linked to failure for the serum Na to rise. He raises the issue that the pathology is not entirely clear, and is only suspected to be due to fluids (perhaps contributory). Do we like this theory only because we can do something about it?

In counterposition to the last speaker, MY points out that many current guidelines still include fluid boluses… (tongue in cheek?)

He then shows, illustrated beautifully, how bad clinicians are at estimating dehydration. pH and bicarbonate aren’t much better! Despite this, MY says, we’re actually doing pretty well at correcting it, and most patients seem to do okay.

Interestingly, studies of bolus therapy in DKA show that slower rehydration appears to resolve acidosis quickier! This is likely to be due to the impact of saline on hyperchloraemia, resulting in acidosis.

Overall, available evidence is pretty thin, but slow resuscitation appears to be the go, over 48 hours or so.

What about bicarbonate? Theoretical benefits, but theoretical risks such as increased ketone production, worsened cerebral oedema and cellular acidosis. A total of just 57 adults in 3 trials suggests there are no clinically important benefits, and possibly some harm.

Does a buffered solution help? Results still unclear – you probably get where you want to go slower with saline, but still get there. As recent literature reinforces, perhaps the increase in chloride associated with normal saline is harmful. For those who don't know, check the Journal Club for Rinaldo Bellomo's recent paper on chloride restriction and renal failure

Todd Fraser wrote 10-25-2012 08:17:48 pm
And that's it - time for a beer!

Todd Fraser wrote 10-26-2012 10:29:26 am
Welcome to day 2

I’ve chosen to go to the update session this morning and see how much I can learn in 90 minutes

Todd Fraser wrote 10-26-2012 10:29:49 am
We start with Mike Anderson, the nomadic intensivist, who runs us through short cuts. Here’s the skinny
• Metanalyses don’t give you much
• Femoral CVC infection risk might be over-stated
• Dobutamine in acute heart failure doesn’t work
• Neither does IABP in cardiogenic shock who undergo PCI

Mike then turns attention to George Skworonski’s paper on the Greying Intensivist. Undoubtedly, older practitioners need to be utilized in different ways, and the current high stress, high on-call practice environment cannot be sustained. Transitioning through this period into retirement is a thorny issue the colleges will need to address.

A final “short cut” – robots in ICU! Run in the US, where there is an estimated deficit of 1200 intensists, a remotely run telerobot has been developed and trialed to allow remote management of patients in a surgical ICU. Perhaps it improves patient safety and communication in the ICU. No, I just can’t see it.

Todd Fraser wrote 10-26-2012 10:30:12 am
Next up is Antony Tobin – what’s new in the thorax?

Tobin begins with the 2010 French paper investigating the role of paralysis in the ventilation management of ARDS. Big doses of cisatracurium were given, no train of 4 measurement was performed, the ARDS-NET protocol was followed, and 90 day hospital mortality was studied. Interestingly, patients with severe disease had a significantly improved outcome. Potential criticisms though – was weakness assessment good enough? It wasn’t blinded, and the mortality reported was adjusted for disease severity (and unadjusted mortality was not significant).

Importantly, ICU myoneuropathy was not more prominent in the intervention group. Tobin reviews the evidence for NMBAs causing weakness and shows it is actually quite weak. He reminds us that spontaneously ventilating patients may generate significant transalveolar pressures – unrecognized unless you are using an oesophageal probe.

Tobin now turns his attention to a 2012 paper by Needham et al looking at the real world impact of a protective lung ventilation strategy on 2 year outcome in patients with ARDS. Interestingly, only 41% of the observations were consistent with the recommendations! Mortality overall was enormous (64% at 2 years), and the authors demonstrate that adherence to the guidelines improved mortality significantly. Interesting.

So what does this mean? Is it just a marker of overall better care? Is it a real effect on the lung, or on inflammatory protein generation? Is this something we should be auditing in our own workplace?

Next he refers to a paper looking at the impact of extubation failure in a paper by Thille in 2011. In this paper, those who failed extubation got sicker, measured by SOFA scores, and had a mortality of 50%, compared with a 5% mortality in those who were successfully extubated.

Following on, he looks at a paper by Desai et al in 2012 studying the methods used to assess the rapid shallow breathing index – T piece, CPAP on 0cmH20 and CPAP on 5cmH20. These were applied to all patients, and the ventilation parameters measured were markedly different in each. This highlights that these modes are NOT interchangeable.

He then reviews a paper by Martin Tobin in 2012 that comments on the “Myth of Minimal Ventilator settings”. The author argues that commonly held belief that some settings were “minimal” were actually still providing significant support, particularly in some patient pathophysiologies, and we should rethink our extubation strategies, possibly including a spontaneous breathing trial.

Tobin concludes by reviewing a couple of papers looking at chloride restriction, culminating in a recent paper by Yunos and Bellomo in JAMA this month. This paper of course showed reduced renal injury in the period that restricted chloride. Another job for the CTG?

Todd Fraser wrote 10-26-2012 10:30:32 am
Michael Reade
Michael eloquently advises us that while there are a number of positive trials, we need to “read the fine print”. The devil is in the detail…

He starts with Tranexamic Acid and the CRASH-2 trial. This positive trial suggested a benefit with TXA when given early in trauma patients at risk of or actually bleeding. All cause mortality was reduced by 2-3%.Some key points – 60% were not shocked. Only 50% required blood transfusion. Yet 16% died! Low rates of thrombembolic diease were reported, but did they go looking? Death rate was higher after 3 hours. But, much like FEAST, this trial was conducted predominantly in developing countries. Does the intervention hold the same risk-benefit profile in the developed world?

He then presents some data from military environment (the Matters trial) that suggests TXA is of benefit – but then demonstrates some problems with this observational trial. He says he feels it is too early to incorporate TXA into trauma resuscitation protocols.

He then reviews an analysis by Ho in 2012 looking at ratios of FFP to red cells in trauma resuscitation. While there appears to be benefit, he points out the significant impact of survival bias on these studies – if you to live long enough to get FFP, you’re probably more likely to survive.

He also refers to some animal studies looking at how long “hypotensive resuscitation” can be maintained. There is a clear point where the animals stopped getting a benefit from this strategy and started dying. We eagerly await the clinical trials!

Changing tack, Michael goes on to a favorite topic of his, delirium in ICU. He starts by looking at the impact of CAM-ICU, a delirium diagnostic tool, on the rate of diagnosis in ICU. The tool seems to be good at picking up cases we would otherwise miss.

He then moves to some pre-trial data looking at early sedation practice in ICUs in Australia. While there is a general belief, he says, that we like to restrict sedation to improve patient outcomes, this is not what we actually practice. Mortality, even after adjustment for severity, appears higher with heavy sedation. All this leads to a new trial to be conducted in Australasia examining this point.

He concludes with the key study looking at no-sedation vrs current practice perfomed by Strom in 2010. He points out that this trial wasn’t “no” sedation, but “no sedation if they don’t need it and you attend properly to their analgesic needs”. There was a substantial trend to a better outcome.

Todd Fraser wrote 10-26-2012 04:44:13 pm
Session 2

I’ve chosen to go to the ARDS update – some interesting speakers here, and looking forward to hearing what’s new.

Todd Fraser wrote 10-26-2012 04:45:36 pm
Andrew Bersten
Reviewing the factors involved, Andrew points to predisposing insults, alcoholism and cigarette smoking as key factors. Interestingly diabetes reduces risk, while increasingly, there is identification of genetics as a factor – more will come in the near future.

Andrew then reviews the pathophysiology – shunt and hypoxia, increased elastance and fibrosis (probably occurring earlier than we realize) are all factors. Interestingly, increased minute ventilation is seen, probably due to worsening pathophysiological deadspace. Pulmonary hypertension, multifactorial in origin, is commonplace.

Two phases are known – infiltrative, later replaced by fibroproliferative. Cell counts are known to rise in BAL fluid, but while neutrophils had previously been implicated, interest is increasing in the role of the alveolar macrophage – ARDS can arise in neutropaenic patients for example.

The initiator of the proinflammatory response is incredibly complex and is evolving all the time. Ultimately, increased permeability leads to proteinaceous oedema. This permeability involves disruption of function of both the endothelium and endothelium. For example, vascular endothelial cadherin dysfunction leads to endothelial leakage. The mechanisms involved in this may be targets for therapy.

Exploring the reasons for lung stiffness, AB points out that the lung tends to collapse normally and is held open by chest wall recoil. However, surface tension is the factor that contributes most to keeping the lung open, and when surface tension is disrupted (by dysfunction of surfactant etc), lung elastance is significantly increased. There are multiple mechanisms that contribute to the surfactant dysfunction including dilution, activity of enzymes and reduced production. Lung injury due to ventilation also contributes, but this is largely only possible if the lung is already diseased – its hard to “cause” VILI in a normal lung.

AB discusses the FACTT trial - why are high fluid therapy arm harmful? Is it just the volume load? AB contends that raised pulmonary capillary pressure, by various mechanisms, is actually pro-inflammatory.

What happens with resolution or ARDS? It requires repair of barrier function, proliferation and replacement of denuded type 1 alveolar cells, clearance of oedema and protein etc. Barrier function needs to be repaired, and pulmonary cellular function is often impaired (clearance of lung water for example). Evidence is accumulating that the fibroproliferative phase starts much earlier than previously recognized, even within 24 hours of intubation.

Questions from the floor
• Antony Tobin – why did beta agonists not work? AB – Perhaps we could look at trying to get the drug there more efficiently, but probably the failure was related to the inability to control the leakiness to start with. Perhaps if we deal with those factors first, the beta agonists may work better. May need to look at this again in the future.

Todd Fraser wrote 10-26-2012 04:46:46 pm
Next session was from Prof Gordon Rubenfeld, talking about current therapies in ARDS.

He starts with the perennial question in ICU – how much PEEP? Everyone agrees on the concept – open the lung, ventilate it gently – but there is almost no agreement on how to set this. A metanalysis of ALVEOLI, LOVS and EXPRESS, all negative high-PEEP trials, showed higher PEEP was of value of benefit in severe disease but harm in less severe (ALI). Interestingly, there was no assessment for likelihood of “PEEP responsiveness” in any of these trials. Sadly this leaves us little closer to consensus, but GR suggests using around 14-15 cmH2O is optimal.

The concept of PEEP responsiveness is novel. They seem to have reduced mortality. Perhaps this is just a marker of less severe disease.

So what do we do at the bedside? Can follow a table like in ARDS-NET – little is better than this at this point in time. GR prioritises plateau pressures ahead of increasing PEEP, and doesn’t increase PEEP to the point that haemodynamic resuscitation is required.

Turning his attention to sedation practice, GR reviews the existing knowledge on sedation protocols – it appears initially they are helpful, but perhaps we’re getting smarter at it anyway. The late mortality benefit seen in the ABC trial is fascinating but inexplicable. Propofol certainly looks better at reducing ventilator duration compared with some benzodiazepines, while dexmeditomidine appears helpful but the studies are still inconclusive. More work is required.

Is sedation required to tolerate protective lung ventilation? A study done by GR suggests that it is possible that it isn’t – but depends on what degree of ventilator asynchrony you’re prepared to tolerate. Perhaps heavy sedation to make the patient “look right” is not required, particularly if there are long term benefits to the patient, and the patient isn’t likely to remember it anyway. He points to the Kress data suggesting that lighter sedation is not associated with higher long term psychological impacts. Also interesting however, is that in that study, despite lower sedation limiting delirium, it did not improve long term psychological and cognitive outcomes.

Pointedly he notes that the most expensive ventilation day is the first day, and no change to sedation protocols will affect this – so cost benefit is less likely to be achievable.

He also suggests that what we need to aim for – mobilizing ventilated patients on inotropes – isn’t feasible in our current configuration. Can we move towards this?

What about paralysis – GR suggests that while 28 day mortality was less, 90 day was not. So, he remains skeptical. However, it appears safer than we thought, and agrees with Antony Tobin that the risks have previously been over-stated.

Question from the floor
• Manoj Sanexa – we tend to use spontaneous breathing strategies a lot – does this matter, if the patient takes too big a breath? GR suggests we simply don’t know – perhaps we’ve recruited and the compliance has improved? Do we sedate to get control? Unclear – bases his management on whether or not he feels that they are improving and nearing extubation.

Todd Fraser wrote 10-26-2012 04:47:11 pm
Final speaker for the session, Andrew Davies from the Alfred in Melbourne, talking on the likely rescue therapy of the future in ARDS.

AD starts by giving a case example – familiar to us all. Do we ventilate, pronate, oscillate or cannulate? Each shows promise, none yet are ahead of the pack.

AD highlights the Amato study – we seem to have forgotten this – is eerily familiar to many of the newer trials in ARDS now. Interestingly, the study had a marked effect on reducing rescue therapies, a trend seen in other open lung strategy trials.

He reviews the Staircase recruitment manoeuvre, to be used in the PHARLAP trial. This has been very effective in the pilot trial.

What about Prone? Gattinoni, he reminds us, found clear benefit in oxygenation, but found no difference in survival. Other trials have found similar results. Perhaps earlier and longer duration is important? The PROSEVA trial, just released, showed 474 patients found prone for 16 hours per day resulted in a halving of 28 and 90 day mortality. Too good to be true? We eagerly await the published trial.

What about HFO? A 2010 review by Sud in the BMJ suggested benefit, but the Oscillate trial in Canada has stopped early – results have not been released, but AD is highly concerned it will be shown to be futile. OSCAR, a trial of 800 patients in the UK, is also due for release in the next few months.

ECMO? This therapy has been done for years. Until recently, the results of most studies was depressingly poor. Improving technology has probably led to better efficiency and reduced complications. The CESAR trial, performed in 2009 by Peek et al, found better outcomes in severe respiratory failure. While small, and limitations exist, this is promising. With vastly increased exposure resulting from the H1N1 epidemic, perhaps ECMO is ready for prime time. The results of the ANZ ECMO work from the pandemic found a 70%+ survival rate, in a cohort that would have expected much much lower survival.

The EOLIA trial of ECMO, a French based study, may answer the role of ECMO in a couple of years. Certainly the technology appears relatively safe, albeit with high levels of very skilled training required.

So overall, stay tuned, because there is more coming. Perhaps the list presented at the top of his talk is a step wise approach.

Todd Fraser wrote 10-27-2012 10:01:52 am
Anders Aneman

We start the afternoon session with a discussion on the starch issue. Clearly the place of starch based resuscitation fluids has been heavily questioned in recent years. Anders presents the findings of a multicentre RCT performed in severe sepsis in Scandanavia. The study allowed for up to 33ml/kg ideal body weight of the trial fluid per day, and was compared with Ringers’ lactate. This paper of course was yet another nail in the coffin of HES, and reinforced the view that HES causes renal failure.

Todd Fraser wrote 10-27-2012 10:02:18 am
Next up in Rinaldo Bellomo, who kindly steps into Simon Finfer’s shoes to present the results of CHEST. Speaking of nails in the coffin…

The basis for this trial was the SAFE TRIPS trial. It found that the predominant fluids used world wide at the beginning of the trial was saline or starch. The lingering doubt over the potential renal toxicity of starches, combined with its widespread use, necessitated the need for a large trial.

CHEST involved 7000 patients requiring fluid resuscitation in ICU, with supportive evidence suggesting hypovolaemia (adopted broadly from SAFE). They used slightly higher doses compared with the scandanavian study, with up to 50ml/kg/day.

Reflecting the findings of SAFE, the ratio of fluid required to achieve similar haemodynamic endpoint variables was not 3:1 as classic physiology teaching would suggest, and the volumes were very similar. Blood was used more often in the starch group.

There was no difference in mortality, while there was a marked worsening of renal failure. Curiously though, lower grades of RIFLE class renal dysfunction were less common in the starch group. This may have been due to the increase in urine output with starch, which limited the number of patients who achieved these lower grades.

Some other interesting tidbits – starch appeared to reduce the need for vasoactive agents, while it increased the likelihood of significant liver impairment. It also increased the risk of pruritis.

No specific subgroup of patients could be identified on an a priori specified analysis could be identified who did better.

Questions from the floor
• Gordon Rubenfeld – is there any remaining argument for giving starch? Both speakers agree that there is not. It seems to be the death knell for starch, particularly with plenty of evidence for this leading into the trials.
• John Green – there was no mortality difference in one trial but was in the other. What was the difference? RB - this is possibly a dose effect, with more given to sicker patients.
• Ian Jenkins – how does this apply in the real world where people give different volumes of these fluids? RB – its always difficult to extrapolate from a controlled, unrealistic environment of a trial to real world, but is a necessary evil to remove bias.
• Therese Jaques – are there starches and starches? Yes, some are potato, some maise, others exist, however, its unclear whether this makes any difference at this point. It remains possible a smaller molecular weight compound could become available that is more easily cleared by the kidney – could this be better tolerated? Possibly…

Todd Fraser wrote 10-27-2012 10:03:35 am
Rinaldo part 2 - this is the talk he was supposed to give!

The master sticks around to chat about chloride restriction.

Its worthwhile noting that the comparator in these trials, and with SAFE, was saline, and perhaps there are issues here. Saline is one of the most commonly used solutions world wide. It is also worth noting saline forms the backbone of albumin and other solutions. All have high chloride content.

Increasingly though are concerns that this may have clinical consequences. The effect on metabolic acidosis is well known. It appears to have effects in renal transplant, on urine output and on other markers of renal function. His most recent trial was designed to attempt to address this.

The trial was a prospective open label before and after trial, where they restricted the use of chloride rich fluids. The vast majority of fluid used was Hartmann’s, with chloride poor albumin increased dramatically too. This represents a drop of chloride administration from an average of 650mmol to 450mmol during their stay. Total fluids administered was similar before and after.

The results suggest a significant reduction in renal injury by either RIFLE, KDIGO or AKIN criteria.

What explains this? Chance? Possibly. Maybe simply time bias as treatments get better. Probably not, as this isn't reflected in changes in the APD. What about avoiding gelofusine? Possibly, as recent evidence (Anaesthetsia and intensive care 2012) suggests, it might be harmful.

What about albumin? SAFE suggests this isn't so. What about 20% albumin? Possibly, no evidence elsewhere for or against this

Many other potential reasons exist - however, there doesn't appear evidence for any of these. D of note, fluid volumes given were similar, and more sodium was given before. Interesting.

This is supported by studies in volunteers that showed a difference in renal blood flow when chloride restricted fluids are given.

Is it time for a trial?

Overall, very suggestive.

Todd Fraser wrote 10-27-2012 10:09:12 am
Kathryn Maitland is back to talk about why fluids might have caused harm in FEAST.


She begins by presenting the video of the expected outcomes from the practitioners involved. They all were convinced that fluid boluses were helping. It highlights the commonly held belief that fluid boluses were essential to resuscitation. Interestingly, the majority of the deaths were in the first 24 hours.


KM addresses the key concerns of the trial :


She refutes the suggestion that the investigators had access to accumulating results, and changed the trial half way through to get a result.


She also clarifies that the children were exceptionally sick, as some authors have suggested otherwise.


The paper was criticized as including patients who were not as sick, by not being socked as per the WHO guidelines (weak pulse reduced capillary refill). The inference is that people are suggesting they weren't sick enough to get a bolus. KM states that meeting this criteria is extremely uncommon, with it failing to include even moderately hypotensive kids, and shows data to suggest that these kids were exceptionally sick. She shows that if you select out the kids who DID meet other definitions of shock, including the WHO's, the mortality is still different. In fact, the mortality difference is MOST profound when you define it as per the WHO's!!!


There was also criticism about the bolus volume, but she demonstrates that the effect was the same before and after the study protocol was changed to increase the volume from 20 to 40ml/kg.


KM divided her study group into three major groups - those presenting with severe respiratory distress, neurological failure and shock, and any combination of these. No matter what group was studied, the results are consistent.


The issue of severe anemia has also been raised, with the theory that bolus therapy in these kids would make it worse, but the evidence suggests that this is not true. There was the same harm shown in this group as all the other groups.


Even the response clinically to the fluids didn't seem to predict a better outcome. They seemed to get better initially, then returned to shock and died quickly after anyway. Of course, there were some who responded quickly and did well.


So why did they die? Fascinatingly, more kids died in the bolus arm died of lethal cardiovascular failure, not worsening respiratory or neurological failure as you might expect. How can this be possible? This data will soon be published, and has wider implications for fluid management.


Perhaps the physiological response to shock is protective in some way. Perhaps there is a repercussion injury scenario. The results are hard to explain.

Todd Fraser wrote 10-27-2012 10:09:43 am
Okay, that's it. Time to go and have a beer with the Panda's at the shindig at Adelaide zoo! See you tomorrow.

Todd Fraser wrote 10-27-2012 01:32:57 pm
Welcome back! I've started the day with an interesting free papers session - here are the highlights from the one I went to.

Todd Fraser wrote 10-27-2012 01:33:30 pm
Contrast enhanced echo – Antoine Schneider

This echo technique utilizes microbubbles injected into the circulation to provide “contrast” during ultrasound. Its been used to assist in echncardiography for some time, but Antoine demonstrates its use for quantification of renal blood flow. Using this technique, they studied renal cortical perfusion in cardiac surgery patients. Sounds like an interesting technique. One problem will be the time element, as scans take about 30 minutes in many patients. There is also quite a lot of variation between observers. Next step is probably a comparison with a gold standard…

Todd Fraser wrote 10-27-2012 01:33:49 pm
Perceived risks, benefits and barriers to SDD implementation – Ian Seppelt

The concept of SDD is reasonably well known, but despite considerable evidence seemingly in support of it, resistance to its implementation is great. The SuDDICU collaborative has attempted to explore the reasons for this, and barriers to conducting the definitive trial. Most clinicians in Australasia are undecided on the impact of SDD, while in the UK and Canada there is higher levels of acceptance. The biggest concern to implementation, particularly in Australasia and Canada is the potential generation of resistance. Despite this, most units appear willing to participate in the definitive RCT.

The plan now is to conduct the definitive effectiveness trial – frighteningly, this is likely to need 30 000 patients!

Todd Fraser wrote 10-27-2012 01:34:16 pm
Augmented renal clearance in Multi-trauma patients – Andrew Udy

Drug dosing is commonly based on pharmacology in normal volunteers, and even when assessed in patients, is rarely studied in ICU. This concept is increasingly recognized in sepsis and burns, but Andrew presents data that this may also be the case in trauma.

He demonstrates a well known phenomenon that creatinine clearance can predict clearance of some important drugs such as antibiotics. Certain antibiotics of course require prolonged time over MIC levels to be effective, so these changes are really important to clinical outcomes.

Andrew then presents details of a single centre observational study looking at trauma patients receiving beta lactams for prophylaxis without raised creatinines, performed early in the patient’s stay, around day 2-3. The average ISS was 29, with 92% requiring mechanical ventilation, so the patients were fairly sick. Interestingly, Creatinine clearance was considered above normal in 86% of them, with an average of 166ml/min/1.73msq. Age appears to be the biggest predictor of this. They also sampled drug concentrations but the data is currently pending.

Todd Fraser wrote 10-27-2012 01:34:35 pm
Hospital mortality of obstetric patients with sepsis – Matthew Maiden

Matt stars by reviewing obstetric physiology, and there are a number of reasons why pregnancy may predispose to sepsis. Sepsis is also the number one cause of maternal death in developed countries. The recent epidemic of influenza had a profound impact on pregnant women too.

Matt analysed the ANZICS adult patient database to assess the impact of pregnancy in critical illness outcomes. Sepsis accounts for a quarter of pregnant patients coming to ICU, and 5% of post partum admissions. By APACHE 2 at least, they were less unwell than non-pregnant women in the same age group. Interestingly, the death rate was significantly lower than predicted by APACHE 2 score. Unfortunately, only 50% of women in this age range had pregnancy status recorded in the database.

How to explain these results? Perhaps quality of data is an issue, while the death rate was so low that it would be difficult to detect a difference. Additionally, APACHE 2 was developed in populations that did not include pregnant women. Perhaps their pregnancy status makes their APACHE 2 score look worse than they really are, overestimating their risk of death. Or alternatively, perhaps to get pregnant in the first place you need to have a reasonable level of health!

Todd Fraser wrote 10-27-2012 01:34:54 pm
Mean Perfusion Pressure During Initial Management of Shock – Rakshit Panwar

It is generally accepted dogma in ICU that patients with background hypertension require higher MAPs during resuscitation, to compensate for altered autoregulation.

The study looked at ICU patients on vasopressor infusions for 72 hours. They looked at fairly sick patients, and estimated the time spent at 60-80% of their estimated basal mean perfusion pressure. Renal injury, as defined by RIFLE criteria, appeared to be related to higher percentages of time spent below acceptable pressures.

Despite widespread awareness of this issue, it was surprising to find that there was no correlation between background blood pressure and the achieved blood pressure – indicating that higher blood pressures were not achieved in patients with hypertension. The reasons for this could not be discerned from this study.

Todd Fraser wrote 10-27-2012 03:18:03 pm
I've chosen the feeding session for this afternoon. Some really big names in this session, should be a cracker

Todd Fraser wrote 10-27-2012 03:19:25 pm
Marianne Chapman – Enteral feeding is un-natural

The answer is no, says Chapman, but the gut continues to conspire against us in ICU. Evidence suggests we still achieve only 80% of feeding goals by day 7 in ICU. The most common reason for this is GIT disturbance, and a smaller contribution from planned withholding for procedures.

What are the reasons for this GIT disturbance? MC begins by describing the problem, best measured by scintography, which is not commonly available outside research. Reviewing examples, she demonstrates that gastric stasis plays a very large part. About half of patients still have >50% of administered feed remaining in the stomach at 4 hours.

The stomach is known to be less distensible, has less peristaltic waves and is poor at co-ordinating movements. This disturbance is worst at the distal stomach. This is accentuated by the pylorus which has increased tone and contractility. Furthermore, the feedback mechanism that shuts off gastric emptying, via the hormone cholecystakinin, caused by feed in the duodenum, is exaggerated – ie the suppression of gastric emptying is amplified for the same amount of feed in the duodenum.

Small intestinal movement is also impaired. MC shows examples where peristalsis in the duodenum may actually be retrograde. This is also seen lower down, with over 50% of peristaltic waves in ICU patients being retrograde. Rauch et al looked at transit time using capsule endoscopy and have found a wide variety of transit times in critical illness, with some very quick and some grossly prolonged.

Back to cholecystakinin – what is its role in all this? Interestingly, the levels are markedly elevated in critical care patients, both fasting and after feed challenges. In addition, in all critically ill patients, CCK levels were significantly higher in those who were feed intolerant. So is it a potential therapeutic target?

Other hormones may be interesting too, such as Gherelin and PYY – more to come in future studies.

There is further evidence of impaired intestinal luminal cell function, resulting in poor absorbtion. What’s the mechanism for this? Reduced enzymatic activity? Reduced blood flow? Mucosal cell atrophy? Other?

Many patients seem to have more than one of these issues, with a compounding result.

Questions from the floor
• Daren Heyland – is there a biomarker of feed intolerance? MC – no, not yet. CCK looks interesting. GRVs are a good enough measure of stomach dysfunction, but we have almost no way of measuring small bowel function clinically.
• John Lambert – does the lack of bolus feeding make a difference? MC – great question, not sure, possibly the dose of feed is insufficient when feeding continuously to “normalize” small bowel function
• David Tuxen – If glucose is poorly absorbed, where does it go? MC – probably into the colon, causing diarrhoea. Can it be measured? Perhaps, we’re looking into it.
• Steve Warrillo – isn’t this just an adaptive response? Should we be trying to overcome nature? MC – yes, possibly, as there is evidence that feed restriction is appropriate as recent studies suggest.

Todd Fraser wrote 10-27-2012 03:20:03 pm
Daren Heyland – EN, what works?

Daren begins by highlighting the evidence that supports EN over PN, but what’s more contentious is the amount. His work in 2011 suggests that a target of 80% appears sound, though this is association not cause-effect. Having said that, the EDEN trial did not show a benefit, as Marianne alluded to. He cautions though that if your BMI is <20 or higher than 35, perhaps this isn’t appropriate – work by Todd Rice in this area suggests this.

Clearly the best dose will needs to be individualized. Nonetheless, earlier feeding is better, and possibly the more you can get in the better.

Failure to achieve this level is almost ubiquitous in ICU worldwide. He stresses that doing the same thing as we’ve done before is unlikely to get different results, hence the PEP UP trial.

In this protocol, they commence at target rate, starting with concentrated semielemental, progressing to polymeric, and use a daily target rather than hourly, which may focus attention on the issue more to ensure we achieve global goals. Promotility agents are used prospectively, not reactively. Protein supplements were given in some patients.

Commencing at goal rate is something fairly unfamiliar to most of us – DH presents data suggesting this approach is just as safe as incremental.

Setting a daily goal rather than hourly prevents the issue of intermittent withholding of feeds. For example, if the feed is paused for a procedure, the rate is increased to compensate for this to ensure the 24 hourly goal is met.

An ongoing concern in feeding is the issue of feeding while on inotropic / vasoconstrictor support. While the evidence is limited by retrospective and observational nature, it appears early feeding in these groups STILL leads to better outcomes.

Do trophic feeds help? At a microscopic level, yes, clearly improves histological appearance of the mucosa. Certainly there appears little reason not to do at least this in nearly every patient, reassessing and escalating as tolerated.

An interesting strategy is to have bedside nurses report nutritional adequacy for the 24 hour period – perhaps this should be a key element of any nursing or medical handover?

PEP UP aimed to review the impact on a protocol on practice (not patient outcome), and clearly it achieved this.

Questions from the floor
• David Tuxen – fluid balance is clearly an issue. What’s you’re view on concentrated feeds? DH – we recommend 1.5cal/ml, but using higher concentrations is associated with its own complications such as diarrhoea
• Steve Warrillow – what are your recommendations for gaining buy-in? DH – insisting on multidisciplinary approach and all taking responsibility. Making it a standard part of any admission to ICU order sets would be helpful
• Geoff Shaw – there is little evidence supporting 80% as a goal over 60%, and given its difficult to get buyin, why not lower it? DH – interesting comment, might be reasonable

Todd Fraser wrote 10-27-2012 03:44:10 pm
Adam Deane – prokinetics

What is best practice for feed intolerance? It appears erythromycin is the best at improving gastric emptying. Metaclopromide is also useful. Unfortunately there are complications, including QT prolongation, enzyme induction, liver effects and the potential for antibiotic resistance.

Erythromycin stimulates the motilin receptor, so are there any other options? GSK962040 is a molecule that seems promising. Another option is Ghrelin. It increases growth hormone secretion and has an effect on gastric emptying.

What about absorbtion? The major concern is that driving feed into the small bowel, and then it not being absorbed, might lead to bacterial overgrowth and other complications. However measuring small bowel absorbtion and doing something about it are next to impossible at this point. Erythromycin appears to improve glucose absorbtion INDEPENDENTLY of its role of gastric emptying, but curiously may reduce lipid absorbtion. Its also worth realizing erythromycin may cause slowing of small bowel transit, so is not indicated in ileus. The impairment of nutrient absorbtion is at least in part due to molecular downregulation of the transport mechanisms, but the reasons for this are still unclear.

Sadly, route of administration (in particular post pyloric) does not appear to assist patient centred outcomes such as mortality or infectious complications, as demonstrated by the ENTERIC trial. He points out that because of the feedback mechanisms, particularly through CCK, gastric emptying is slowed even during small bowel feeding, resulting in retention of gastric secretions. This may be why ENTERIC showed no difference in VAP reduction.

Questions from the floor
• Steve Warrillow – is reduced absorbtion an adaptive response, so is augmenting it appropriate? AD – ventilation isn’t natural either, so its not that straightforward. The issue is more likely to be what the target should be.

Todd Fraser wrote 10-27-2012 03:45:29 pm
So that's it for the scientific program, and only the ANZICS oration remains. Who will it be?

Jonno Reade wrote 11-26-2012 09:29:10 pm
Hi guys,

I just wanted to let you know that I really enjoyed the blog from the conference this year. I couldn't make it but it was terrific to hear what was going on. I've also read what you wrote for the CICM ASM earlier in the year. Nice work.

Is this another #FOAMed initiative?

Todd Fraser wrote 11-28-2012 08:07:00 am
Thanks Jonno, that's nice of you. I enjoyed doing it, and it makes it all the more worthwhile hearing that you did too.

I think #FOAMed is Free Open Access Meducation. Its a great concept that is trying to bring together medical educators who are providing free education on line.

Our free content includes this blog, all our podcasts (which are available on iTunes) and some of our videos on YouTube. The iPad app is pretty cheap too!

;-)