Steroids (CS) in TBI

CRASH (2004) Corticosteroid Randomisation After Significant Head Injury

Randomised placebo controlled study multicentre trial (239 hospitals, 49 countries excl USA)
10,008 pts recruited (initially planned 20,000 but trial stopped early)
Inclusion: head injury + GCS 14 within 8hrs of injury
Intervention: Methylpred 2g loading + 0.4gm/hr infusion x48hrs vs placebo
Background:
o Physiological plausibility: Steroids (CS) thought to reduce intracranial oedema + pressure, improve
cerebral blood flow and reduce occurrence of herniation thus improve outcome
o Beneficial effects seen in cerebral oedema 2 brain tumours/surgery and acute spinal cord injury
o Systematic review (1997) suggested CS reduce risk of death 1-2%
1 outcomes: death within 2weeks + death/disability at 6months
Results:
o CS group had sig. higher all-cause mortality @ 2weeks (21.1% vs. 17.9%, p=0.0001)
o 6month mortality higher in CS group (25.7% vs 22.3%, p=0.001) + trend toward increased
combined endpoint of death or severe disability (38.1% vs 36.3%, p=0.08)
Comments:
o Methodology: large, international study. 96.7% 6month follow up. Early randomisation (<8hrs)
o Few details re: other management pre or post randomisation
o No reporting of incidence of concomitant spinal cord injury
o Unclear mechanism of increased mortality with CS
Similar rate of complications (seizures, GI bleeding, infection) in both groups

Recommendation: Steroids should not be used routinely in treatment of acute traumatic brain injury based
on results of CRASH.

Therapeutic Hypothermia (TH) in TBI

Eurotherm 3235 (2015) Hypothermia for Intracranial Hypertension after Traumatic Brain Injury

Randomised open label international multicentre (neuro ICU centres) study (47 hospitals, 18 countries)
387 pts recruited (initially planned 600 but trial stopped early)
Inclusion: Persistent ICP>20 despite stage 1 treatments (MV, sedation), closed TBI with abnormal CTB, initial
head injury <10 days earlier
Intervention: Hypothermia (32-35) fot at least 48hrs + standard care vs standard care alone
Background:
o Physiological plausibility: Fever in TBI detrimental. Decrease in CRMO 2 with temp (7%/C)
1 outcomes: GOS-E @ 6months. 2 outcomes: 6month mortality, lack of ICP control, incidence pneumonia,
ICU LOS, MOHS @ 28days
Results:
o TH group GOS-E scores shifted in unfavourable direction (adj OR 1.53 [CI 1.02-2.03 p+0.04]
o Higher risk of death in TH group (hazard ratio 1.45 [CI 1.01-2.1 p=0.047]
Comments:
o Supports other higher quality studies which show trends towards unfavourable outcomes
o Stage 3 interventions required more often (barbiturates) or earlier (decompressive craniectomy) in
control group suggesting that TH does reduce ICP (or at least the need for other stage 2
therapies)
o The Eurotherm3235 Trial provides evidence against therapeutic hypothermia as an early measure
to lower intracranial pressure but no guidance on its use for refractory intracranial hypertension,
for which its benefits may plausibly outweigh risks
Recommendation:
o TH in current practice could still be considered as a rescue therapy for refractory ICP after
analgesia, sedation, osmotic therapy, ventriculostomy and NMB.
o Await results of POLAR-RCT (TH as early intervention (pre-hospital/ED) for TBI)
EPO-TBI (2015, Nichol, international)

- In patients with moderate or severe traumatic brain injury does the administration of erythropoietin
compared with placebo improve neurological outcome at 6 months after injury
- MCRCT, partial double blinding, 29 centres, n=606
- Intervention: EPO weekly up to 3 doses vs placebo
- Outcome:
o No difference in 6 month GOS-E
o No difference in proximal DVT or composite thrombotic outcomes
- Conclusion: Following moderate or severe TBI, EPO did not reduce no. of pts with severe neurological
dysfunction or increase the incidence of DVT or composite thrombotic outcomes
- Criticisms:
o Study powered to detect RRR 24% - cannot exclude smaller risk reduction
- Summary:
o This trial does not support the use of Erythropoietin for patients with moderate or severe TBI.
o It does alert the neurointensivist to the high incidence of proximal DVT (16-18%) in this
patient group, although the clinical significance of this is not explored in this trial
Decompressive Craniectomy (DC) in TBI

DECRA (2011) Early Decompressive Craniectomy in Patients with Severe Traumatic Brain Injury

MCPRCT (neuro ICU centres) study (15 centres in Aus, NZ, Saudi Arabia), n=155
Inclusion: Severe TBI (GCS 3-8 or Marshall grade III on CT). ICP >20mmHg for >15mins in 1hr despite
optimisation of 1st tier (sedation, normal CO2, osmotherapies, CSF drainage, NMB)
Intervention: Earlier decompressive craniectomy (DC) vs standard care
Background: - Physiological plausibility:
1 outcomes: GOS-E @ 6months. 2 outcomes: Time of ICP, ICP index, ICU/hosp LOS, in-hospital and
6month mortality
Results:
o ICPs better controlled in DC group. Shorter duration MV and ICU/Hosp LOS but more surgical
complications
o Trend towards shifting survivors to unfavourable outcomes (composite death, vegetative state,
severe disability) in 70% DC group vs 50% control
?axonal stretch or disruption of brain architecture as proposed mechanism
Comments:
o In control group 19/82 (23.2%) underwent DC (4 early, 15 later)
o Study relates to 1 particular surgical approach only (bifrontal craniectomy)
o 1 centre responsible for 30% of pts
Recommendation/In my practice:
o DC may be a useful option when maximal medical treatment has failed to control ICP
o Potential benefit in paediatric populations
o Awaiting RESCUEicp (DC as last resort vs as an early/primary intervention)

RescueICP (2016, Hutchinson)

- Pts with TBI and refractory intracranial hypertension, does decompressive craniectomy, result in more
favourable mortality and neurological outcomes at 6 months, compared with barbiturate coma and
continued medical management?
- MCRCT, non-blinded, international (70% from UK), n=408
- Inclusion: TBI with abnormal CTB, ICP>25 for 1-12hrs despite stage 1 and 2 measures
- Decompressive craniectomy (unilateral FTP or bifrontal) vs barbiturate therapy
o Cont stage 1 (positioning, sedation, mech vent, paralysis and stage 2 (EVD, osmotherapy,
hypothermia, loop diuretics)
- Results: 37.2% of medical group later received DC. 9.4% of surgical group received barbiturates
- Outcomes:
o If GOS-E (6months) is grouped into 3 categories Dead, VS/LSD and USD or better
craniectomy group: 26.9%, 30.3%, 42.8%
medical group: 48.9%, 16.5%, 34.6%
- Conclusion: At 6 months, DC in TBI with refractory ICP resulted in mortality + higher rates of
vegetative state, lower severe disability, and upper severe disability than medical care
- Criticisms:
o Slow recruitment over 10yrs
o Use of TH as stage 2 therapy may have had deleterious effect (based on Eurotherm 3235)
o A large proportion of patients in the medical group underwent decompressive craniectomy
(37%); this situation may have diluted the observed treatment effect
- Summary:
o DC after stage 1-2 measures was associated with mortality however, more survivors in DC
group were dependent on others.
o Concern (from DECRA + RescueICP) that life-saving surgery may not predictably result in
sufficiently good functional survival. Further investigations, exploring patient selection, longer
term recovery and quality of life may be indicated
Complications of DC:
- Infection, bleeding, seizures, hygroma, collections (subgaleal/subdural)
- Brain herniation through craniotomy, venous thrombosis from herniation through defect
- Sinking flap syndrome (Paradoxical subtentorial herniation with LP or CSF drainage (due to atmospheric
pressure))
- Hydrocephalus
- Worsening of brain injury
Decompressive Hemicraniectomy (DHC) in Malignant MCA Stroke

Rationale:
MCA infarction has a substantial mortality and morbidity.
Factors which produce this effect include:
o Large volume of infarcted brain tissue, leading to significantly increased ICP.
o Significant risk for haemorrhagic transformation.
o Significant midline shift, with pressure on medial cerebral structures .
o Temporal herniation, with compression of the posterior cerebral artery.
o Poor perfusion of the contralateral cerebral hemisphere due to raised intracranial pressure.
Many of these factors are related to the increased intracranial pressure.
DHC can decrease intracranial pressure by increasing cranial compliance.
Thus DHC should be able to improve mortality and morbidity from acute MMCAS by vastly improving
the perfusion in the penumbra of the stroke shortly after the craniectomy.

DESTINY trial (2007):

Prospective, multicenter RCT in Germany
32 pts (initially planned 188pts) but stopped early due to statistically significant mortality reduction
Raw data suggests improved survival for the craniectomy group: 88% vs 47%.
DECIMAL trial (2007):
Prospective, multicenter RCT in France
After randomization of 38 patients (18-55), the data safety monitoring committee recommended
stopping the trial because of slow recruitment.
Absolute reduction of 52.8% in the death rate in the surgery group.
Survival was 22% in the "conservative management" group, and ~ 75% in the surgery group
HAMLET trial (2009):
Prospective, multicenter RCT in the Netherlands
64 patients were randomised
Survival was better with surgery (absolute risk reduction was 38%)

Pooled analysis of the European studies

A pooled analysis the three studies (93pt cases): probability of survival increases from 28% to nearly 80% and
the probability of survival with an modified Rankin Scale of 3 doubles.

DESTINY II Trial (2014): DC for MMCAS in older people

112 patients, older than 60 years (median age was 70)
1 outcome: survival without severe disability; 38% vs 18% in the control group.
2 outcome: lower mortality 33% vs 70% in the control group.
o However, in contrast to young pts - vast majority of the post-operative survivors were
severely disabled.
PATCH (2016, Baharoglu, Europe)

- In patients with acute intracerebral haemorrhage (ICH), associated with antiplatelet therapy, does
platelet transfusion compared with standard care reduce death or dependence?
- MCRCT, open label (assessors blinded), 41 centres, n=190
- Inclusion: Non-traumatic ICH with GCS 8-15 on antiplatelets without intraventricular extension or
imminent surgery
- Intervention: Platelet transfusion (1x for COX inhibitor, 2x for ADP receptor inhibitor)
- Outcome:
o Adjusted OR 2.05 for shift towards death / dependence at 3 months in intervention group
o Non significant signal of benefit in subgroup with ICH volume >30mL
- Conclusion: Platelet transfusions cannot be recommended in acute ICH in pts on antiplatelet therapy
- Criticisms:
o Differences in baseline characteristics.
- Summary:
o This small RCT shows plt transfusion leads to worse neurological outcomes at 3 months

ATACH-2 (2016, Qureshi, multinational)

- In patients with acute ICH with hypertension, does rapid lowering of SBP improve patient outcomes?
- MCRCT, open label, 110 centres, n=1000
- Inclusion: Non-traumatic ICH with volume <60ml, within 4.5hrs of symptoms
- Intervention: Target SBP 110-140 for 24hrs vs control SBP 140-180 for 24hrs with IV CCB infusion
- Outcome:
o No difference in mRS 4-6 (significant disability or death) at 3 months: 38.7% vs 37.7%
o No difference in proportion with ICH extension >33% at 24hrs
- Conclusion: The results do not support reduction of SBP to 110-140mmHg in ICH
- Criticisms:
o Underpowered (calculated on event rate of 60% vs observed 38%)
o More than 50% pts recruited from Asia sites ?generalisability
o Significant treatment failure (primary and secondary) in treatment group
- Summary:
o The results of this trial and INTERACT2 trial do not support an early, intensive control of SBP in
patients with acute ICH

INTERACT-2 (2015, Anderson, multinational)

- In patients with acute ICH with hypertension, does intensive lowering of SBP improve patient outcomes?
- MCRCT, open label, outcome assessor blinded, 144 centres, n=2839
- Inclusion: Non-traumatic ICH, within 6hrs of symptoms
- Intervention: Target SBP <140 for 7 days vs control SBP <180 for 7 days
- Outcome:
o No difference in mRS 4-6 (significant disability or death) at 3 months: 52.0% vs 55.6%
o Better functional outcome with lower mRS in intervention group
- Conclusion: Intensive lowering of blood pressure did not significantly reduce rate of primary outcome
of death or severe disability. An ordinal analysis of modified Rankin scores indicated improved
functional outcomes with intensive lowering of blood pressure
- Criticisms:
o No standardisation of anti-HTN used
o More than 60% pts recruited from Asia sites ?generalisability
- Summary:
o In patients with intracerebral haemorrhage intensive blood pressure lowering did not reduce
the risk of death or severe disability.
ISAT (2002, Molyneux, Europe/USA)

- In patients with intracranial aneurysms, is endovascular coiling a safer alternative to clipping?

- MCRCT, open label, 43 centres, n=2143
- Inclusion: Aneurysmal SAH amenable to either coiling or clipping
- Intervention: Endovascular coiling vs craniotomy + clipping
- Outcome:
o mRS 3-6 at 2 months: 25.4% vs 36.4%, at 1 year: 23.7% vs 30.6% (RRR 22.6%)
o Risk of rebleeding at 1 year: 2/1276 vs 0/1081. Coiling required more further procedures
- Conclusion: In aneurysmal SAH amenable to either endovascular coiling or neurosurgical clipping,
survival free of disability at 1 year is significantly better with endovascular coiling.
- Criticisms:
o Lack of angiographic data following the initial treatment and long-term follow-up
o Concerns regarding expertise bias regarding patient selection and center participation criteria
- Summary:
o In patients with a ruptured intracranial aneurysm, for which endovascular coiling and
neurosurgical clipping are therapeutic options, the outcome in terms of survival free of
disability at 1 year is significantly better with endovascular coiling.
EGDT Literature

Comparison of EGDT studies

Rivers et al ProMISe ProCESS ARISE
Location US UK US Australasia
Population 263 1260 1351 1600
Sepsis Definition
Suspected / Actual Yes Yes Yes Yes
Infection
SIRS criteria 2 Yes Yes Yes Yes
Refractory BP or Yes Yes Yes Yes
lactate > 4 mmol/l
Protocol
Fluid before 2030 ml/kg 1000 ml ~2030 ml/kg 1000 ml
randomisation Changed during study
Antibiotics before Yes No Yes
randomisation
Recruitment not specified <6h from ED arrival <12h from ED arrival <6h from ED arrival &
& <2h from shock & <2h from shock <2h from shock
criteria criteria criteria
Intervention EGDT 6 EGDT 6 hours EGDT 6 hours EGDT 6 hours
hours
Control Usual Usual therapy 1) Protocol usual Usual therapy
therapy therapy
2) Usual therapy
Primary outcome In-hospital 90-day mortality 60-day mortality 90-day mortality
mortality
Primary Outcome
Intervention 30.5% n/a 21.0% 18.6%
Control 46.5% n/a 1) 18.2% 18.8%
2) 18.9%

EGDT Protocol:
- Art line + CVC capable of ScvO2 monitoring
- Supplemental O2 to Sats >93%
- 500mls bolus of crystalloid or colloid at least every 30 minutes until CVP > 8mmHg (self-ventilating) or >
12mmHg (NIV or invasive ventilation)
- Vasopressors to achieve MAP of 65 and/or SBP 90mmHg
- ScVO2 >70% once CVP and MAP targets achieved
- If ScVO2 < 70% and HCT < 30% PRBC transfusion
- If ScVO2 remains < 70% despite HCT >30% or Hb > 100g/dl dobutamine 2.520mcg/kg/min
- If ScVO2 still < 70% increase oxygen NIV Mechanical ventilation

ARISE (2014)
- MCRCT unblinded, 51 centres in Australasia, 1600 pts
- Intervention: 6hrs EGDT by team trained in EGDT vs usual resuscitation care
- Outcome: 1: No difference 90-day mortality 18.6% EGDT vs. 18.8% usual care group (RR 0.98, p < 0.9)
o 2: No difference in hospital or ICU LOS, need and duration of organ support, vasopressor
requirement - EGDT 76.3% vs. usual care group 65.8%, P<0.001
- Criticisms/Comments:
o Lower APACHE scores than ProCESS and Rivers study - addressed by subgroup analysis of
APACHE II < 25 vs > 25 No difference in mortality between EGDT and usual-care in the
sicker group although the total numbers were small in > 25 group (n=69)

ProCESS (2014)
 - MCRCT unblinded, 31 centres in US, 1351 pts
- Intervention: 6hrs EGDT with dedicated staff vs protocol based standard therapy group with dedicated
staff (expert opinion) vs usual care group
- Outcome: 1: No difference 60-day mortality EGDT 21.0%; Protocol-based standard therapy 18.2%;
Usual care 18.9%. (P values 0.31 0.89)
o 2: No difference in 90-day or 1 yr mortality
- Criticisms/Comments:
o Changed inclusion criteria during trial (reduced fluid bolus required before meeting
"refractory hypotension" criteria), but mean volume used was within Rivers' original definition
of 2030 ml/kg.
o Mortality ~20% but initial power calculation based on 3046%, therefore interim adjustment
made and recruitment target reduced.
o Adherence to protocol was 88.1% in EGDT group and 95.6% in protocol-based standard
therapy group. Although pragmatic, this is not perfect and may reduce between group
differences.

ProMISe (2015)
- MCRCT unblended, 56 centres in UK, 1260 pts
- Intervention: 6hrs EGDT with trained staff vs usual care:
- Outcome: No difference 90-day mortality EGDT 29.5% vs 29.2% (p=0.9)
o 2: No difference in other outcomes except slightly lower ICU LOS and SOFA scores in control
group
- Comments/Criticisms:
o Important outcome measures incl. cost analysis + QALY
o Study population representative of critically ill septic shock pts with APACHE II 18 (Arise 15)
o Good adherence to EGDT 86-95%
o Only 1/3rd eligible patients recruited

PRISM (2017)
- Meta-analysis of ProMISe, ProCESS and ARISE
- 90-day mortality was statistically identical for EGDT (24.9%) and usual care (25.4%)
- EGDT was associated with more ICU days (mean 5.3 days vs. 4.9 days) and vasoactive drug use (1.9 days
vs. 1.6 days)
- There were higher costs with EGDT
- No benefit from EGDT for patients with worse septic shock (higher lactate levels, predicted mortality, or
with combined hypotension and hyperlactatemia)
- PRISM's sickest one-third of patients by risk scoring had a 45% mortality in both the EGDT and usual
care arms, and was 4x the size of the entire 2001 EGDT trial.
SEPSISPAM (2014, Asfar, France)

- In patients with septic shock does a high target mean arterial pressure (MAP), compared to a low target
MAP, improve mortality?
- MCRCT, blinding to pts and research staff, 29 centres, n=776
- Inclusion: Septic shock pts refractory to fluids with vasopressors <6hrs
- Intervention: Target MAP 80-85 vs 65-70 (observed values usually 70-75)
- Outcome:
o No significant difference in 28day mortality: 36.6% vs 34%
o More new AF in high MAP group: 6.7% vs 2.8%
o Less need for RRT in high MAP group: 31.7% vs 42.2% p=0.046
o Chronic HTN sub-group doubling Cr less in high MAP group: 38.9% vs 52% p=0.02
- Conclusion: Targeting a MAP of 80-85, compared with 65-70 did not affect mortality
- Criticisms:
o Underpowered due to lower than expected mortality
o (Frequent) use of steroids and aPC may limit generalisability
o Achieved MAP higher than target MAP
- Summary:
o Majority of septic shock pts a target MAP of 65-70 is adequate.
o Consider a higher MAP in chronic HTN pts as reduced RRT with a NNT of 9.5
o In patients without hypertension further studies comparing a MAP of 65 with a lower MAP of
e.g. 55 would be beneficial

ALBIOS (2014, Caironi, Italy)

- In adults with severe sepsis or septic shock, does 20% albumin solution with crystalloid fluid compared to
crystalloid fluid alone reduce death with 28 days?
- MCRCT, non-blinded, 100 centres, n=1818
- Inclusion: Severe sepsis pts without head injury, CCF, or specific indication for albumin
- Intervention: 20% albumin 300mL daily to maintain serum albumin >30g/L vs crystalloid only at
clinicians discretion
- Outcome:
o No difference in 28 days (and 90 day) mortality: 31.8% vs 32%
- Conclusion: The addition of albumin to crystalloids during the first 28 days of treatment to maintain a
serum albumin level of 30 g/l or more is safe, but does not provide a survival advantage over
crystalloids alone over a follow-up period of 90 days
- Criticisms:
o Open-label, non-blinded methodology may have influenced clinicians' decisions, which may
have biased toward either intervention or control.
o Lower incidence of death compared to anticipated rate (45% used in power calculation) may
to type 2 error (failure to detect a true positive outcome)
- Summary:
o The use of 20% albumin in adults with severe sepsis will improve haemodynamic indices, but
will not reduce mortality. Its use is safe but not recommended for routine fluid resuscitation.
Targeting a serum albumin of 30g/l or more does not appear to have a survival advantage.

SAFE (2014, Australasia)

- Does fluid resuscitation with albumin, compared to saline, affect mortality for patients in the ICU?
- MCRCT, double blinded, 16 centres, n=6997
- Inclusion: ICU pts requiring IVF excluding CTSx pts, liver tx, burns
- Intervention: Resuscitation with 4% albumin vs 0.9% Saline
- Outcome:
o No difference in all-cause mortality at 28 days: 20.9% vs 21.1%
o Decreased mortality in severe sepsis 30.7% vs 35.3% (prompting ALBIOS)
o Increased mortality in trauma 13.6% vs 10% (prompting SAFE-TBI)
o SAFE-TBI: Increased mortality in head injury: 24.5% vs 15.1% p=0.009
- Conclusion: Use of 4% albumin or 0.9% sodium chloride for fluid resuscitation results in similar
outcomes at 28 days
- Criticisms:
o Minimal weaknesses. The authors note that the sub-group analyses had insufficient power.
- Summary:
o Overall there is no difference when 4% albumin is used for fluid resuscitation when compared
to 0.9% sodium chloride.
o 4% albumin should be avoided in head injuries.
Catecholamines and Fluids

VASST (2008) trial:

- Vasopressin vs noradrenaline in septic shock

- MCRCT double blinded, 26 centres 778 pts
- Intervention: Vasopressin vs. Noradrenaline (Doses of open-label vasopressors given according to
clinical indication)
- Outcome: 28-day mortality no significant difference: VP 35.4% vs NA 39.3%
o Subgroup of those with less severe shock appeared to benefit.
o No detrimental effect in terms of cardiac safety in septic shock pts
o No difference in 90-day mortality, need for CRRT, vasopressors, steroids, LOS
- Comments/criticisms:
o Majority of pts received open label norad with study drug hence study really examines
addition of vasopressin to standard vasopressors vs standard alone
o Low dose vasopressin (max dose 1.8u/hr)
- Conclusion:
o No definite evidence that vasopressin as an addition or replacement to standard vasopressors
has mortality benefit (when MAP adequately maintained)
o Continue using vasopressin as a catecholamine sparing agent where appropriate or in
vasopressor refractory septic shock

VANISH (2016)

- Effect of Early Vasopressin vs Norepinephrine on Kidney Failure in Patients With Septic Shock
- Factorial (2x2) MCRCT, double blinded, 18 centres, 409 pts
- Outcome: No effect on rates of biochemical renal failure: 43% vs 40.8%
o Dialysis rates lower in vasopressin group: 25.4% vs 35.3% (mostly non survivors)
o Vasopressor adverse events marginally higher 10.7% vs 8.3% (mesenteric or digit ischaemia)
- Conclusion: cannot recommend vasopressin as replacement for noradrenaline as first line agent

CATS (2007, Annane, France)

- In patients with septic shock, is mortality affected by the use of adrenaline (epinephrine) alone in
comparison to noradrenaline dobutamine?
- MCRCT, triple blinded, 19 centres, n=330 (1591 assessed)
- Inclusion: Septic shock pts requiring fluids and vasopressors
- Intervention: Adrenaline vs Norad + dobutamine (when needed)
- Outcome:
o No difference in all cause mortality at 28days: 40% vs 34% p=0.31
o No difference in rate of severe arrhythmias, myocardial events, LOS, duration of pressors
- Conclusion: Physicians could use either adrenaline alone or noradrenaline dobutamine in patients
with septic shock
- Criticisms:
o Number of patients excluded - 98 due to physician refusal and 409 due to 'other reasons'
- Summary:
o No significant difference found between adrenaline vs. norad dobutamine in septic shock.
o However, was under-powered and had a trend towards harm in the adrenaline group
o The study does provide some reassurance that adrenaline infusion is a reasonable alternative
in the early phase of shock or in resource limited settings.

SPLIT (2015, Young, NZ)

- In critically ill patients, does the use of a balanced crystalloid solution compared to normal saline effect
the incidence of acute kidney injury?
- MCRCT, double-blinded, 4 centres, cluster double cross over, n=2278
- Inclusion: all ICU pts requiring crystalloid fluid therapy as clinically indicated without ESRF/RRT
- Intervention: Plasmalyte 148 vs NSaline. Rate and frequency determined by treating clinician
- Outcomes:
o No difference in incidence of AKI (RIFLE injury based on Cr) 9.6% vs 9.2% p=0.77
o No difference in incidence of AKI or change in Cr (RIFLE and KDIGO definitions)
o No difference in need for RRT, mech vent, ICU/hosp LOS, mortality
- Conclusion: Among patients receiving crystalloid fluid therapy in the ICU, use of a buffered crystalloid
compared with 0.9% sodium chloride did not reduce the risk of AKI
- Criticisms:
o High proportion of post elective surgery 56% vs 58%, high proportion cardiac surgery (~50%)
o >90% of pts exposed to IVF before enrolment (majority was buffered crystalloid)
o Potential for ascertainment bias as 2/3rd clinicians able to identify fluid type (from
hyperchloraemic acidosis)
- Summary:
o No difference in incidence of AKI in pts receiving Plasma-Lyte 148 compared to 0.9% saline.
o Whilst more data from the trial is awaited, I will continue to use balanced solutions.
o The study does provide reassurance that in elective or moderately sick ICU pts, giving up to 2L
NSaline results in no increased risk of kidney injury compared with Plasma-Lyte 148
Steroids in Meningitis

Evidence: Cochrane review (2013) by Brouwer et al; 25 trials, n=4121.

- The mortality benefit was statistically non-significant
- Deafness was reduced (RR 0.67) as were neurological sequelae of all sorts (RR 0.83).
- Mortality improved in the Strep .pneumoniae patients only
- In children, protection against hearing loss was only seen in the H.influenzae group.
- In third world countries, there was no benefit for anybody, except for those with tuberculosis
meningitis.
- Subgroup analysis of high quality studies did not show any effect on severe hearing loss.
- There was an increased risk of recurrent fever, but no other side effects.

Steroids in CAP

Meta-analysis (Siemieniuk, 2015)

- 3% decrease in all cause mortality
- 5% decrease in rate of mechanical ventilation
- Reduced hospital LOS 1 day
- Critique:
o Included trials small
o High heterogeneity
o Insufficient power to analyse mortality
o Many exclusions eg immunocompromised pts, pregnancy, GI bleeding, neuropsychiatric
conditions prone to psychotic effects steroids
o Small overall effect

Pancreatitis

Steroids in Sepsis

HYPRESS (2016, Keh, Germany)

- In patients with severe sepsis does hydrocortisone compared to placebo prevent the development of
septic shock?
- MCRCT, double-blinded, Germany, n=380
- Inclusion: Pts with severe sepsis without shock
- Hydrocortisone 50mg bolus + infusion 200mg/day x5 days then weaning dose vs placebo (mannitol)
- Outcome:
o No difference in occurrence of shock 21.2% vs 22.9% overall or in subgroups
o No difference with 2 outcomes mortality, LOS, mech vent.
o 206pts had short synacthen testing 33.5% had CIRCI (Critical Illness Related Corticosteroid
Insufficiency). No difference in primary or secondary endpoints between patients with or
without CIRCI
- Conclusion: Administration of hydrocortisone did not prevent the development of shock in patients
with severe sepsis.
- Criticisms:
o Relatively less sick cohort overall mortality rate 8.5%
o Only some sites tested baseline adrenal function
o Early septic shock may have been missed due to need for consent prior to randomisation
- Summary:
o No support for role of routine corticosteroids in severe sepsis
o Await results of ADRENAL
IRONMAN (2016, Litton et al, Perth Australia)

- In critically ill patients who are anaemic, does early administration of intravenous iron compared to
placebo reduce the requirement for blood transfusion?
- RCT, double blinded, 4 centres n=140
- ICU pts with Hb <100 without sepsis
- IV iron carboxymaltose vs saline with repeat dosing from day 4
- Outcome:
o No significant difference in no. PRBC transfusion 1 vs 1
o Median Hb at discharge otherwise no difference in mortality, LOS, adverse events
- Conclusion: In anaemic ICU pts IV iron does not reduce PRBC transfusion, however they did leave
hospital with a higher haemoglobin level.
- Criticisms:
o Limited generalizability large proportion of CTSx and trauma pts (only 14% medical)
o Mean PRBC t/f 1.9 vs expected 4.0 (underpowered possible type 2 error)
- Summary:
o Point estimates suggestive that IV iron beneficial but study inconclusive due to unanticipated
low transfusion rate.
o Further studies required

ABLE (2015, Lacroix, Canada/Europe)

- In critically ill patients requiring blood, does transfusion with fresh RBCs (stored for 8 days or less)
compared with standard issue RBCs (stored 2-42 days) reduce 90 day mortality?
- MCRCT, double blinded, 64 centres in Canada, UK, Europe. n=2510
- Intervention: Fresh blood <8 days or freshest available vs standard therapy
- Outcome:
o No difference in 90 day mortality: 37% vs 35%
o No difference in other mortality indices, duration organ supports, LOS, transfusion reactions
- Conclusion: Transfusion of fresh red cells, as compared with standard-issue red cells, did not decrease
the 90-day mortality among critically ill patients
- Criticisms:
o Lack of uniform transfusion guidelines or protocols however was a pragmatic study
- Summary:
o This study shows that fresh blood < 8 days is no better than blood stored up to 42 days.
o Transfusion triggers and processes for safe administration are probably more important
HEAT (2015, Young, ANZICS)
- Does the regular administration of paracetamol to critically ill patients with fever and known or
suspected infection, affect the number of ICU-free days?
- MCRCT, double blinded, 23 centres, n=700
- Inclusion: T>38C + suspected infection + ABx
- Intervention: Paracetamol 1g QID vs placebo
- Outcome:
o No difference in median ICU free days to D28: 23 vs 22 p=0.07
o No difference in all cause mortality, organ support free days
o Decreased LOS in survivors, increased LOS in non-survivors
- Conclusion: Early administration of paracetamol to treat fever due to probable infection does not
affect the number of ICU-free days
- Criticisms:
o ICU free days a composite outcome of mortality and ICU LOS
o High frequency of protocol deviations: paracetamol group 30%, placebo 28%
o No information regarding paracetamol use pre randomisation or post ICU discharge
- Summary:
o Early administration of paracetamol to ICU pts to treat fever due to probable infection results
in a moderate reduction in temperature but does not affect mortality at 28 or 90 days.
o The observation that it may result in a longer ICU stay in nonsurvivors and shorter ICU stay in
survivors should be interpreted cautiously and only as hypothesis generating
OXYGEN-ICU (2016, Giradis, Italy)

- In critically ill adults, does conservative oxygen therapy compared to liberal oxygen therapy reduce
mortality?
- RCT non blinded, single centre. n=434
- Inclusion: ICU pts with LOS >72hrs
- Conservative O2 therapy Sats 94-98% (PO2 70-100) vs standard 97-100% (PO2 <150)
o Supplemental O2 if Sats <94% for intubation, suction, hospital transfer
- Results:
o Median Fio2 0.36 vs 0.39 (p<0.001), Median PO2 87 vs 102 (p<0.001)
- Outcome:
o ICU mortality: 11.6% vs 20.2% (ARR 8.6% p=0.01)
o NNT 12, Fragility index 3
o Hospital mortality, ventilator free hours
- Conclusion: For critically ill patients with an ICU length of stay of >72 hours, a conservative vs. a
conventional protocol for oxygen therapy resulted in a lower ICU mortality
- Criticisms:
o Single centre, non-blinded
o Unplanned early termination 2 to earthquake (effect overestimation) underpowered
o Low fragility index
- Summary:
o Suggests conservative O2 strategy in critical ill may result in dramatic mortality benefit.
o Further multi-centre studies required (Single centre under powered study)

Helmets Vs Facemask NIV (2016, Patel, USA)

- In patients with acute respiratory distress syndrome (ARDS), does the use of non-invasive ventilation (NIV)
delivered by helmet compared to NIV delivered by face mask reduce the need for intubation?
- RCT, non-blinded, single centre, n=83 (stopped early for efficacy)
- Inclusion: Pts with ARDS (Berlin criteria) requiring NIV for at least 8hrs
- NIV delivered by helmet (ICU ventilator) vs facemask (non-invasive ventilator)
- Outcome:
o intubation rate 18.2% vs 61.5% p<0.001, NNT 3, fragility index 10. ARR 43.3%
o vent free days, ICU LOS, hosp mortality, 90 day mortality
o No difference hosp LOS, mask related pressure areas
- Conclusion: In this single-center, randomised clinical trial, NIV delivered by helmet significantly reduced
the intubation rate amongst patients with ARDS compared with the patients receiving NIV by face mask.
The helmet was also associated with improved ventilator-free days and significantly reduced ICU length
of stay as well as 90-day mortality
- Criticisms:
o Likely cohort with mild ARDS (need for 8hrs NIV likely excluded more hypoxic pts)
Median duration of NIV after randomisation 19.8 vs. 26.4 hours
o Multiple protocol changes to broaden inclusion criteria to include all causes of resp failure
o Early termination of the trial likely to overestimate the treatment effect
- Summary:
o Helmet NIV shows promise as a tool to reduce the need for intubation in patients with rapidly
reversible hypoxaemic respiratory failure. However, larger trials are needed to determine if
these results are generalisable outside of this small single-centre study

HOT-ER (2016, Jones, NZ)

- In hypoxic adults presenting to the emergency department, does nasal high-flow oxygen compared to
standard oxygen therapy reduce the need for advanced respiratory support?
- Single centre RCT, non-blinded, n=322
- Inclusion: ED pts with Sats <92% + RR>22 who did not require immediate NIV
- HFNP (starting at 40L/min, 37C Fio2 0.28) vs standard O2 therapy (NP, VM, HM)
o Protocols for conversion to NIV or intubation according to British Thoracic Society Guidelines)
 - Outcome:
o No difference in rate of NIV or IPPV 3.6% vs 7.3%. Fragility index 3
o Failure to tolerate HFNP 1 in 12 (8.5%)
- Conclusion: HFNP did not reduce the need for NIV or IPPV vs standard oxygen therapy
- Criticisms:
o Non-blinded with subjective primary outcome at clinicians discretion
o Failed to recruit target number, so under-powered to detect the expected difference
o Significant loss-to-follow-up for patient experience survey strong likelihood of bias in
reporting
- Summary:
o This trial did not demonstrate a difference between HFNP and standard O2 therapy in hypoxic
and dyspnoeic ED patients
o Significant weaknesses in the methodology mean that no firm conclusion can be drawn, as a
false negative result may have occurred

Post-extubation HFNP (2016, Hernandez, Spain)

- In patients that are low risk for re-intubation does high flow nasal oxygen compared with conventional
oxygen therapy reduce re-intubation within 72 hours?
- MCRCT, non-blinded to clinicians, 7 centres, n=527 (out of 10347)
- Inclusion: Mech vent pts with low risk for reintubation scheduled for extubation
o Age<65, BMI<30, few comorbidities, absence of: heart failure, mod-severe COPD, airway
patency issues
- Intervention: HFNPx24hrs post extubation vs conventional therapy- NP or NRBM to Sats>92%
- Outcome:
o Significant reduction in reintubation @72hrs 4.9% vs 12.2%, p=0.004, fragility index 5, NNT 14
o No difference in ICU/hosp mortality, resp infections
o Significantly lower post extubation resp failure in HFNP group
o Result still significant even after excluding reintubation for laryngeal oedema
- Conclusion: In patients that were low risk for requiring re-intubation the use of high flow nasal oxygen
reduced the risk of re-intubation
- Criticisms:
o Large number of excluded patients (>80% who were not weanable) limits external validity.
o Reasonably high reintubation rate in control group of low-risk pts. May limit external validity
of centres with a lower re-intubation rate in low risk patients.
o Some baseline differences between groups eg significantly lower neurological co-morbidity in
the high flow nasal oxygen group. Potential confounder.
- Summary:
o This MCRCT found that HFNP in low risk pts reduced the risk of re-intubation at 72 hours
o However, due to some baseline differences, and some concerns over the external validity I
would want to see further trials before this becomes routine practice.

FLORALI (2015, Frat, France)

- In pts with acute hypoxaemic respiratory failure, does HFNP vs facemask or NIV prevent intubation?
- MCRCT, open-label, 23 centres, n=310
- Inclusion: Acute T1RF without chronic resp disease, APO, haemodynamic instability or vasopressor use
- Intervention: HFNP OR NIV (8hrs/day with HFNP in interim) vs standard facemask, for 2 calendar days
- Outcome:
o No difference in intubation rate @28days : 38% (HFNP), 50% (NIV), 47% (FM) p=0.18
o Benefit with HFNP on ICU and 90 day mortality.
o In subgroup with PF<200, benefit with HFNP on intubation rate, ICU and 90 day mortality
- Conclusion: In pts with T1RF, treatment with HFNP, standard face mask oxygen, or NIV did not result in
a significantly different intubation rates. There was a significant difference in favour of high-flow nasal
oxygen in 90 day mortality
- Criticisms:
 o Significant protocol violation with NIV group receiving few actual hrs (<8hrs) of NIV and given
FM instead with significant crossover with standard facemask group
o Majority of pts had pneumonia (82%)
- Summary:
o This study demonstrates HFNP is non-inferior to oxygen delivered by a non-rebreather
facemask and BiPAP in reducing the subsequent need for intubation.
o HFNP reduces ICU and 90 day mortality compared with the other strategies. It also
subjectively improves dyspnoea and respiratory discomfort at one hour compared with the
other oxygen delivery devices.
o Further studies are required to confirm whether high flow oxygen is beneficial in pulmonary
and extra pulmonary causes of hypoxaemia and to determine its efficacy in ARDS.

Preoxygenation and apnoeic oxygenation

PREOXYFLOW (2015, Vourch)

- Compared to HFFM (high flow face mask), HFNC as a preoxygenation device did not reduce the lowest
level of desaturation in an RCT
THRIVE (2015, Pateal)
- Case series of 25 patients with difficult airways undergoing general anaesthesia for hypopharyngeal or
laryngotracheal surgery had mean apnoea times of 14 minutes without desaturation (i.e. SaO2 >90%)

FELLOW (2015, Semler, USA)

- Does apneic oxygenation during laryngoscopy in ICU pts increase the lowest arterial oxygen saturation
experienced by patients
- Single centre RCT, single-blinding, n=150 (196 eligible)
- Inclusion: ICU pts being intubated by a fellow, excluded if: intubation too urgent, thought to require
AFOI or VL)
- Intervention: NP @ 15L/min during induction and intubation. Extra O2 via NRBM, NIV, standard NP
- Outcome:
o No difference in lowest SaO2 between induction and 2mins post successful intubation
o No significant difference in incidence of hypoxaemia, desaturation or tertiary outcomes
- Conclusion:
o Apnoeic oxygenation does not appear to increase lowest arterial oxygen saturation during
endotracheal intubation of critically ill patients compared to usual care
- Criticisms:
o A small trial with a surrogate primary outcome. A much larger trial would be required to
detect a clinical outcome such as hypoxemic arrest in the peri-intubation period
o Pts excluded were perhaps sicker or deemed difficult intubations where apnoeic
oxygenation would be most useful.
o Intubators were fellows (presumably skilled) this intervention cannot be discounted if there
are less skilled operators
o Single centred, medical ICU limiting external validity
o The lag in SpO2 may be >2mins following intubation in some cases
o The median highest FiO2 in the preceding 6 hours was 40%. Apneic oxygenation would be
expected to assist if the patient had a high FiO2 at the time of intubation and this may explain
why a treatment effect has not been observed in this study as profoundly hypoxemic patients
are underrepresented.
- Summary:
o In this medical ICU, where patients were being intubated by experienced fellows, and where
the operator did not think that video laryngoscopy was essential, apnoeic oxygenation did not
increase lowest arterial oxygen saturation compared to usual care. No harm was
demonstrated.
NIVAS (2016, Jaber, France)

- In pts who develop acute hypoxaemic respiratory failure post abdo surgery, does NIV prevent
reintubation?
- MCRCT, open-label, 20 centres, n=293
- Inclusion: Acute T1RF post abdo surgery
- Intervention: NIV (encouraged at least 6hrs/day) vs standard O2 therapy (no HFNP allowed)
- Outcome:
o Decreased reintubation within 7 days: 33% (NIV) vs 45% p=0.03, NNT 9. Fragility 2
o Decreased pneumonia and nosocomial infection in NIV
- Conclusion: In pts with established T1RF after abdominal surgery, NIV reduces the risk of re-intubation
within 7 days compared to standard oxygen therapy.
- Criticisms:
o Lower than expected reintubation rate
o Need for reintubation is subjective (potential for bias from unblended clinicians)
o Low fragility index
- Summary:
o 6hrs of NIV reduced re-intubation in T1RF post abdo surgery
o Significant result but relatively fragile; no clinically relevant harm or intolerance was identified
and therefore this paper will change my practice
o Further studies comparing this strategy to high-flow nasal oxygen therapy are urgently
needed

NIVAS (2016, Jaber, France)

- In pts who develop acute hypoxaemic respiratory failure post abdo surgery, does NIV prevent
reintubation?
- MCRCT, open-label, 20 centres, n=293
- Inclusion: Acute T1RF post abdo surgery
- Intervention: NIV (encouraged at least 6hrs/day) vs standard O2 therapy (no HFNP allowed)
- Outcome:
o Decreased reintubation within 7 days: 33% (NIV) vs 45% p=0.03, NNT 9. Fragility 2
o Decreased pneumonia and nosocomial infection in NIV
- Conclusion: In pts with established T1RF after abdominal surgery, NIV reduces the risk of re-intubation
within 7 days compared to standard oxygen therapy.
- Criticisms:
o Lower than expected reintubation rate
o Need for reintubation is subjective (potential for bias from unblended clinicians)
o Low fragility index
- Summary:
o 6hrs of NIV reduced re-intubation in T1RF post abdo surgery
o Significant result but relatively fragile; no clinically relevant harm or intolerance was identified
and therefore this paper will change my practice
o Further studies comparing this strategy to high-flow nasal oxygen therapy are urgently
needed
Heparin Citrate Study (2015, Gattas, ANZ)

- In critically ill pts requiring CRRT, does regional citrate prolong circuit life compared to heparin protamine?
- MCRCT. Single blinding of statistician. 7 centres, n=212 (857 circuits)
- Intervention: Regional citrate + normocalcaemia vs regional heparin protamine anticoagulation
- Outcome:
o Functional circuit life longer in citrate group
o Median circuit life of 1st circuit 39hrs vs 22.8hrs
o Median circuit life of clotted circuits 16.5hrs vs 11.8hrs
o No difference in ICU LOS, mortality, IL-6, IL-8, IL-10, PRBC t/f, duration of CRRT
o More adverse effects in heparin group 11 vs 2 (mostly suspected or confirmed HITS)
- Conclusion: Regional citrate prolongs CRRT circuit life compared with heparin-protamine
anticoagulation, does not affect cytokine levels, and is associated with fewer adverse events
- Criticisms:
o More pts admitted to ICU from ED in heparin group (35.5% vs 22.9%)
o Unblinded
o Underpowered to detect patient-centered outcomes such as mortality, time in ICU, time in
hospital and renal recovery
o A cost analysis would have been useful
o Large no. of pts excluded from study limits external validity 1219 patients excluded
- Summary:
o Regional anticoagulation with citrate and calcium extends filter life in comparison with
regional heparin and systemic reversal with protamine
o Given the greater efficiency and safety of citrate, serious consideration should be given to its
use as first line anticoagulation in continuous renal replacement.
PermiT (2015, Arabi, Canada + Saudi)

- In critically ill adults, does restriction of non-protein calories (permissive underfeeding) compared to
standard feeding reduce mortality at 90 days?
- MCRCT, non-blinded but objective endpoints, 7 centres, n=894 (14% of screened pts)
- Inclusion: EN within 48hrs ICU admission, excluding high dose vasopressors, pregnancy, burns, liver
transplant, condition with <6months life expectancy, post cardiac arrest
- Intervention: Permissive underfeeding 40-60% vs 70-100% caloric requirement
o Caloric requirements calculated with Penn State (BMI<30) or Ireton-Jones (BMI>30 or spont
breathing), protein goal 1.2-1.5g/kg/day
- Outcomes:
o No statistically significant difference in 90 day mortality 27.2% vs 28.9%
o No difference in various mortality rates, serial SOFA scores
o Post hoc: lower incidence of RRT in underfed group (7.1% vs 11.4%)
- Conclusion: Moderate caloric feeding with maintenance of full protein requirement, compared to
standard full feeding, was not associated with reduced mortality
- Criticisms:
o 86% of screened patients were not randomised, which may limit external validity
o Non blinded and clinician led management of non-nutritional care may introduce bias
o Only notable outcome is incidence of RRT but should be interpreted cautiously (post hoc)
- Summary:
o This trial has not demonstrated a moderate survival benefit from permissive underfeeding
with moderate caloric intake (around 50% of target calories)
o A small survival benefit may exist but this study was not large enough to detect one
o Permissive underfeeding with full protein requirement appears safe in critically ill patients

PYTHON (2015, Bekker, Netherlands)

- In patients with severe pancreatitis, does early enteral feeding compared with on-demand feeding
reduce death or major infection?
- MCRCT, blinding of outcome assesors only, 19 centres, n=208 (867 screened)
- Inclusion: Predicted Severe acute pancreatitis excluding recurrent/chronic pancreatitis or 2 ERCP and
malignancy.
- Intervention: Early feeding via NJ tube within 24hrs, feeds gradually up titrated from 20ml/hr vs on-
demand group oral diet from 72hrs unless requested by pt earlier, if unable to tolerate NJ feeds
- Outcomes:
o No difference in composite death or major infection within 6months: 30% vs 27%
o Slow time to full oral diet in early group: 9 days vs 6 days p=0.001
o Only 31% of on demand group required NJ tube placement
o No difference in necrotising pancreatitis, ICU admission,
- Conclusion:
o Early enteral feeding in these patients did not improve outcomes when compared to delayed
on-demand feeding at 72 hours
o This result disagrees with current guidelines which advise feeding to provide a protective
effect on intestinal integrity and systemic nutritional status
- Criticisms:
o Scoring systems poor at predicting eventual severity so cases of non-severe pancreatitis may
have been included
o Unblinded treating clinicians
o NJ tubes used, when nasogastric tubes are easier to place and may be as effective
o Not able to inform management of other forms of pancreatitis
- Summary:
o Safe to allow pts with severe pancreatitis 3-4 days to initiate oral intake
o Although well-conducted this was a relatively small trial, and future data may change this
picture
CALORIES (2015, Harvey, UK)

- In critically ill patients, does parenteral nutrition compared with enteral nutrition improve mortality?
- MCRCT, non-blinded, 33 centres, n=2400
- Inclusion: Non elective ICU pts expected to need nutritional support >48hrs
- Intervention: PN vs EN via NG/NJ with goal 25kcal/kg/day.
- Outcome:
o No difference in all-cause mortality at 90 days: 33.1% vs 34.2%
o Less vomiting and clinically significant BSL in PN group but serious BSL not different
o No difference in duration organ support, SOFA scores, ICU/hosp LOS, mortality, infectious and
non-infectious complications (aspiration events or pneumonia)
- Conclusion: Early nutritional support through the parenteral route is neither more harmful nor more
beneficial than such support through the enteral route
- Criticisms:
o Non blinded study however pragmatic trial with primary endpoint not subject to bias
o Cross over occurred in 6.8% PN group, 3.4% EN group
o Caloric targets not met in majority of both groups although similar (21.3kCal/kg vs 18.5kCal/kg)
- Summary:
o Early nutritional support with PN is neither more harmful nor more beneficial than EN
o Enteral feeding does increase vomiting and hypoglycaemia but without evidence of harm or
nosocomial infection.
o Daily calorific targets were rarely achieved in this study (< 40% in both groups). This reflects a
possible deficiency in our feeding protocols

Early PN when EN contraindicated (2014, Doig, Australasia)

- In critically ill patients with relative contraindications to early enteral nutrition (EN), does providing early
parenteral nutrition (PN), improve mortality?
- MCRCT, single blinded, 31 centres, n=1372
- Inclusion: ICU pts ineligible for EN due to relatively short term contraindication
- Intervention: Early PN (<24hrs of admission) with Harris Benedict equation vs standard care (not
defined by protocol)
- Outcome:
o No difference in 60 day mortality: 21.5% vs 22.8%
o Shorter duration mech vent in Early group: 7.4 vs 6.3 days
- Conclusion: Early PN did not result in significant differences in 60 day mortality or infection rates
- Criticisms:
o Non-blinding of study intervention and outcome assessment
o ~40% of standard care group received PN, the majority within a few days. Therefore it is not
surprising that there was no difference in infectious complications found
- Summary:
o In critically ill patients with a short-term relative contraindication to enteral nutrition, the use
of early parenteral nutrition did not affect 60 day mortality or infection rates compared with
starting feeding at ~3 days.

EPaNIC (2011)

- Early aggressive PN glucose loading via low protein PN vs late PN. Pts subject to tight glycaemic control
- Better outcomes in late group
- In a post hoc subgroup analysis of the EPaNIC trial, 517 patients who were admitted to the ICU with a
surgical contraindication for enteral feeding had fewer infections and an increased likelihood of earlier
live discharge from the ICU if they were not assigned to receive early parenteral nutrition.
- For these patients, starvation was tolerated for 1 week in the ICU and resulted in improved clinical
outcomes.14 Thus, the most effective time at which the initiation of parenteral nutrition can produce a
clear clinical benefit during critical illness remains unclear.
EDEN, Rice, 2011

- Trophic feeds (25%) vs full feeds in ALI/ARDS for 6 days. 1000 pts (who were not malnourished at
baseline)
- No mortality difference between groups
- More GIT intolerance in full feed group (vomit, high GRV, constipation)

Supplemental PN (SPN)

how to treat patients who are eligible to be fed enterally but who cannot tolerate full
enteral feeding after 3 days (12% of admissions in the SPN study
305 pts
Supplemental PN if not achieving 60% of goals determined by indirect calorimetry from day
4-8
No difference in rate of infection in first 28 days
No other secondary outcomes affected

REGANE

329 pts
Gastric residual volumes up to 500mL safe
Allowed more enteral feeding but no change in outcome

NUTRIREA 1

449 pts
Omission of measuring gastric residual volumes did not increase incidence of aspiration or related
complications
Allowed more enteral feeding but no change in outcome

OMEGA

No benefit and possibly harm with omega 3 fatty acids

REDOX and SIGNET

No benefit and possibly harm respectively of glutamine supplementation

Tracman (2014, Young, UK)

- In mechanically ventilated adult patients with a high risk of prolonged ventilation, does early
tracheostomy compared with late tracheostomy reduce mortality at 30 days?
- MCRCT, non-blinded, 70 centres, n=909 (3147 assessed)
- Inclusion: Intubated pts at day 4 where clinician feels pt likely to need at least 7 days
- Intervention: Trache within 4 days vs day 10 or later if still indicated
- Results:
o Early group: 84.6% trache as planned, 6.8% no trache (died, too unstable or recovered), 7.7%
late (too unwell, no resources, trial error)
o Late group: 93.6% planned protocol, 40% trache after day 10, 53.7% not trache as no longer
indicated. 7.3% received trache <day 10 (clinical decision, insistence of relatives, trial error)
- Outcome:
o No difference in all-cause mortality at 30 days: 30.8% vs 31.5%
o No difference in survival at ICU discharge, hospital discharge, 1-year and 2-year follow-up
o Median length of ICU stay were the same: early 13.0 days vs late 13.1 days
- Conclusion: In mechanically ventilated adult patients, there is no mortality benefit from performing an
early tracheostomy
- Criticism:
o Did not include patients requiring tracheostomy for non-respiratory reasons, such as
neurological disease
o Did not look at long-term complications and morbidity associated with tracheostomy
o Inclusion was reliant on clinicians identifying patients at high risk of prolonged ventilation,
with no validated scoring system to assist decision making. This study shows that 19.6%
recovered before day 10 in late group, suggesting that clinician judgement is not good for this.
- Summary:
o This trial does not provide evidence supporting early tracheostomies performed around day 4
compared to day 10
o Tracheostomy procedures are associated with a 6.3% incidence of complications requiring
interventions, which must be weighed against any potential benefit.
o Many commentators have interpreted this trial to mean there is no benefit from performing
early tracheostomies (evidence of no difference), as many will not need it if we wait.
o However, some patients need tracheostomies for good reasons (e.g. neurological disease) and
they were not included in this study.

Cochrane review by Andriolo et al (2015):

- n=1977, of which 909 were from TracMan.

- There was a statistically significant mortality benefit associated with early tracheostomy (47% vs 53%);
the NNT was 11. Mortality benefit was seen at the longest reported follow-up, rather than at 30 days.
- Data regarding risk of pneumonia could not be subjected to meta-analysis due to heterogeneity.
- Decreased duration of sedation and shorter ICU stay but no difference in duration of mechanical
ventilation
- (of note: non-Cochrane meta-analysis by Szakamany 2015 did not replicate mortality benefit)
SLEAP (2012, Mehta, USA/Canada)

- In patients who are mechanically ventilated and sedated with opioids and / or benzodiazepines, do daily
sedative interruptions reduce the duration until extubation?
- MCRCT, non-blinded, 16 centres, n=430 (2091 screened)
- Inclusion: Expected >48hrs mech. ventilation. No cardiac arrest, TBI, NMBD
- Intervention: Opioid and BZD infusions stopped daily restarted at half rate when clinician deemed
necessary vs no planned interruption + titration to RASS/SAS. Propofol, ketamine, dexmedetomidine
not allowed
- Outcome:
o No difference in time to extubation
o Nursing workload increased in intervention group
- Conclusion: When nurses implement a sedation protocol that targets light sedation, the use of daily
sedative interruptions does not reduce the duration until extubation and may increase both the
amount of sedative administered and the nursing workload.
- Criticisms:
o Poor adherence to protocol (85% of all eligible pt days)
o Did not allow propofol which is commonly used over benzodiazepines in ANZ context
o High baseline reintubation and tracheostomy rate. ?limited generalisability
- Summary:
o If a sedation protocol is used to target light sedation within an ICU, daily interruptions are
not shown to be beneficial, are very hard to implement consistently and may increase nursing
workload

ABC (2008, Girard, USA)

- In mechanically ventilated and sedated patients, do Spontaneous Awakening Trials (SAT) in addition to
Spontaneous Breathing Trials (SBT) reduce dependence on mechanical ventilation?
- MCRCT, non-blinded, 4 centres, n=383 (1658 screened)
- Inclusion: Expected MV >12hrs, excluding MV>2weeks, cardiac arrest, profound neuro deficits
- Intervention: Spont awakening trial (de-sedated for 4hrs) progressing to SBT vs SBT alone. Sedation
targeted to level of arousal and comfort deemed appropriate for each pt
- Outcome:
o Increase in vent free days (VFD) 14.7 vs 11.6 p=0.02
o Decreased LOS ICU, hospital. 1yr mortality 44% vs 55%. Trend to 28 day mortality.
o Significantly more self extubation 10% vs 4%
o No difference in reintubation or tracheostomy rate or duration of delirium
- Conclusion: use of a so-called wake up and breathe protocol that pairs daily spontaneous awakening
trials (ie, interruption of sedatives) with daily spontaneous breathing trials for the management of
mechanically ventilated patients in intensive care results in better outcomes than current standard
approaches and should become routine practice.
- Criticisms:
o Inherent biases physician refusal to enrol. Subjective treatment targets comfort
o Complicated statistical analysis due to unusual distribution of the variable
o Mechanism of observed survival benefit difficult to explain
o No surgical pts included due to continual need for analgesia
o Sedation practices have changed since publication
- Summary:
o SATs in this trial appears to expedite weaning from mech vent
o But the results may potentially reflect greater oversedation in control group or bias towards
early extubation in the intervention group.
o Their use may be associated with greater risk of self-extubation and increased nursing
workload
CRASH 2 (2010, International)

- In trauma patients with or at risk of significant haemorrhage, does the early administration of a short
course of tranexamic acid (TXA) affect the mortality, incidence of occlusive events and the amount of
blood transfused?
- MCRCT, double blinded, 274 centres in 40 countries, n=20207
- Inclusion: Trauma pts with significant haemorrhage or at risk (SBP<90 +/- HR>110), clinician unsure
whether to treat with TXA
- Intervention: TXA 1g loading + 1g over 8hrs vs placebo
- Outcome:
o 28 day mortality benefit: 14.5% vs 16%, RR 0.91, ARR 1.5%, p=0.0035. NNT 68
o Death by bleeding: 4.9% vs 5.7%, p=0.0077. NNT 115
o No difference in blood transfusion, surgical intervention, vascular occlusive events
- Conclusion: TXA safely reduced the risk of death in bleeding trauma patients in this study. On the basis
of these results, TXA should be considered for use in bleeding trauma patients.
- Criticisms:
o Randomisation based on subjective opinion of treating physician; no standardised criteria for
the definite use of TXA selection bias
o No stratification of injury severity
o Follow up period only 28 days and was incomplete
o Only 5% pts actually died of bleeding in both groups
- Summary:
o Small absolute risk reduction in mortality with the use of TXA in trauma patients.
o No reduction in amount of blood products administered.
o However, TXA is unlikely to cause harm and hence will continue to be part of practice in the
management of the bleeding trauma patient
 NEURO

NASICS 3 (1997)
ISAT (2002) aSAH: Coiling (vs clipping) improves survival free of disability
INTERACT-2 (2015) ICH: intensive SBP control <140 did not mort or severe disability
ATACH-2 (2016) ICH: intensive SBP control <140 did not mort or severe disability
PATCH (2016) ICH: platelet transfusion for APT harmful death/dependence
DESTINY (2007) MMCAS: DHC improves survival 88% vs 47%
DECIMAL (2007) MMCAS: DHC improves survival 75% vs 22%
HAMLET (2009) MMCAS: DHC improves survival ARR 38%
DESTINY II (2014) MMCAS: DHC in >60yo; improves mortality but many survivors severely
disabled
DECRA (2011) Early DC in TBI: Shifts survivors to unfavourable outcomes; but improves ICP
RescueICP (2016) DC in TBI: Improved mortality but survivors more dependent. Unclear if
improves good functional survival. Confounded by use of TH
EPO-TBI (2015) EPO in severe TBI did not improve functional outcomes
Eurotherm 3235 (2015) TH in ICP (TBI) reduces need for other ICP therapies but leads to worse
functional outcomes + mortality
POLAR-RCT Pre-hospital TH for TBI
CRASH (2004) Steroids in TBI mortality by unclear mechanism

TRAUMA

CRASH2 (2011) TXA: Small mortality benefit ARR 1.5%, death by bleeding ARR 0.8%; greater if
given early <3hrs; no difference in VTE
MATTERS (Military) TXA: Retro cohort: Mortality benefit + decr PRBC but more DVT
MATTERS II (Military) TXA: Retro cohort: Mortality benefit 50%

SEDATION + SBT

SLEAP (2012, Mehta) No difference in time to extubation; more sedatives used in intervention
group; did not use propofol.
ABC (2008) Spont awakening trials + SBT expedites weaning in USA
Esteban (1999) SBT of 30mins equivalent to 120mins
Yang and Tobin (1999) RSBI (RR/Vt) >105 100% sens for extubation failure
Zhang (2014) RSBI >75 threshold when doing SBT with PS 5-7, PEEP 5
DahLIA (2016) Dexmed may be beneficial in ventilated pts with agitated delirium with
improved vent free hours
MIDEX-PRODEX (2014) Dexmed non-inferior to prop/midaz for long-term sedation in ICU

RESP

ACURASYS (Papazian 2010) NMBx in ARDS improves mortality ARR 10%, RRR 25%
PROSEVA (2013) Proning in ARDS improves mortality ~50% if done 16hrs/day
OSCAR (2013) HFOV in ARDS improves O2 but may worsen mortality
OSCILLATE (2013) HFOV in ARDS improves O2 but may worsen mortality
Wan M-A (2004) Massive PE: rTPA improves death+recurrent PE 19%9%
TOPCOAT (2014) Submassive PE: Unclear mortality benefit but improves HD/functional
outcomes
PEITHO (2014) Submassive PE: Unclear mortality benefit but improves HD/functional
outcomes
MOPPETT (2013) Large PE clot burden: Safe dose TPA leads to less chronic pHTN
Cochrane Review NIV in cardiac APO: survival, intubation
Cochrane Review NIV in COPD: mortality 50% vs intubation
Carrillo (2012) NIV in OSA/OHS: Chronic maintenance and rescue in acute resp failure
Chiumillo (2013) NIV in rib #s: Improves mortality, intubation, pneumonia
Lim (2012) NIV in asthma: intubation, ICU LOS, delivery nebs. No mort benefit
Hilbert (2001) NIV in mixed immunosuppression: intubation, in hosp mort
Antonelli (2000) NIV in solid organ transplant: intubation, ICU mortality
Lemiale (2015) NIV in immunocompromised: No difference vs standard care
NIVAS (2016) NIV reduced re-intubation (fragile) in T1RF post abdo surgery
THRIVE HFNP: Apnoeic oxygenation feasible in procedural sedation
Cochrane (2015, Andriolo) Trache: Early trache improves long-term mortality ARR 6%, NNT11
Szakamany M-A (2015) Trache: Did not replicate mortality benefit of early trache
TRACMAN (2014) Trache: No benefit for d4 vs d10 trache in pts with resp disease
FLORALI (2015) HFNP in PF<200 intubation, ICU/90day mortality
Hernandez (2016) HFNP post extubation reduces reintubation at 72hrs
HOT-ER (2016) No benefit of HFNP in hypoxic/dyspnoeic all comers in ED. Poor quality
OXYGEN-ICU (2016) Conservative O2 strategy (94-98%) may reduce mortality

NUTRITION

PermiT (2015) Permissive underfeeding with full protein appears safe in critical care
PYTHON (2015) SAP: safe to allow acute pancreatitis 3-4 days to initiate enteral nutrition
CALORIES (2015) Early nutritional support with PN vs EN: no difference
Early PN when EN C/I (2014, No benefit or harm of early PN when EN C/I in short term vs starting feeds at 3
Doig) days
EPaNIC (2011) Early PN worse than late PN (with glucose loading) ?generalisability
EDEN (2011) Trophic feeds (~25%) safe in ARDS (1st week) in non-malnourished pts
REGANE (2010) Gastric residual vol up to 500mL safe; no change in outcomes
NUTRIREA-1 (2014) Not measuring GRV did not increase complications or change outcome
OMEGA No benefit and possibly harm with omega 3 fatty acids
REDOX/SIGNET No benefit and possibly harm with glutamine supplementation

MISCELLANEOUS

HEAT (2015) Paracetamol to treat fever from ?infection does not affect ICU-free days
ABLE (2015) Fresh PRBC (<8 days) does not impact outcomes
IRONMAN (2016) IV iron may Hb but does not affect transfusion requirements
Gattas (2015) Citrate anticoagulation extends filter life in CRRT vs heparin. Underpowered to
detect pt centred outcomes

SEPSIS/FLUIDS

HYPRESS (2016) Steroids does not prevent septic shock in severe sepsis
Siemieniuk M-A (2015) Steroids in CAP mortality 3%; small overall effect; poor quality
Cochrane Review (2013) Steroids in meningitis: mortality in strep p, deafness h.inf in paeds

SOAP (2006) Excess fluid resus correlates with mortality OR 1.1/1L within 72hrs
SPLIT (2015) Plasmalyte vs NaCl: no difference in AKI when using <2L

CATS (2007) No outcome difference for adrenaline vs NAd dobutamine

VANISH (2016) VP vs NAd in AKI: no difference in AKI but less RRT in VP
VASST (2008) No definite evidence that VP as addition/replacement to NAd has mortality
benefit (when MAP adequately maintained)
SAFE (2014) 4% albumin vs saline in sepsis: No difference all cause mortality
Improved mortality in severe sepsis (prompting ALBIOS)
Increased mortality in TBI (SAFE-TBI)
ALBIOS (2014) 20% alb vs crystalloid in sepsis: HD but not mortality. Safe
No benefit of targeting serum alb >30g/L
SEPSISPAM (2014) MAP 65-70 adequate in majority of septic shock. Less AKI with MAP
PRISM (2017) M-A of ProMISe, ProCESS, ARISE no mortality difference with EGDT vs usual
care. costs + ICU days with EGDT

CARDIO

CHEETAH (2017, Landoni) No benefit of prophylactic levo in CTSx in periop LV dysfunction pts
Levo-CTS (2016, Mehta) No benefit of prophylactic levo in CTSx in pts with LVEF<35%
LeoPARDS (2016) No difference in organ dysfunction in severe sepsis
RUSSLAN Levo in acute LVF post MI improves mortality
LIDO Levo in acute/decomp CHF improves HD but not necessarily mortality
SURVIVE Levo in acute decomp heart failure: no improvement in 6month mort.