1. 20 October 2014 Rachael Pery-Johnston
JOURNAL CLUB RBWH ICU
20 October 2014
Rachael Pery-Johnston

2. The paper:
NHMRC / The Alfred Foundation funding

3. Background to this study:

4. A quick refresher…

5. EGDT: Rivers et al 2001 Clinical Question
In adult patients with severe sepsis or septic shock, does early goal-directed therapy before admission to ICU reduce in- hospital mortality, multiorgan dysfunction and use of hospital resources?
Design
Randomised, controlled, partially blinded trial
263 randomized to 6 hours of either:
Standard Rx (133) or
EGDT (130)
236/263 completed 6 hour initial study (90%)

6. EGDT: Rivers et al 2001 Setting
Consecutive eligible pts in one academic US ED March 1997 to March 2000
Population
Inclusion: adults arriving in the emergency department (ED) with
Suspected infection
2 or more SIRS criteria
Refractory hypotension (BP<90 after 20-30ml/kg fluid)
or hypoperfusion (lactate >/= 4)
Baseline characteristics similar between groups

7. Protocol for early goal-directed therapy

8. Outcomes: Rivers

9. This lead to: Surviving Sepsis Campaign

10. Current guidelines:

11. ProCESS May 2014 Clinical Question
In adult patients with sepsis, does protocol-based care compared to usual care reduce death within 60 days?
Design
Randomised, controlled trial of 3 groups in 1:1:1 ratio
Protocolised EGDT vs
Protocolised standard care vs
Standard care
Non-blinded design
Setting
31 academic hospitals in USA
March 2008 to May 2013

12. ProCESS May 2014 Population
1,351 randomised within 12 hours of arrival in ED
Inclusion: adults arriving in the emergency department (ED) with:
suspected sepsis
refractory hypotension or lactate > 4 mmol/l
two or more SIRS criteria
Exclusion: if another primary diagnosis was present, such as acute myocardial infarction or CVA, and

13. ProCESS May 2014 Intervention Group
Early goal-directed therapy (EGDT) group: strict protocolised care (based on Rivers study) for 6 hours
Dedicated doctor, nurse and research assistant that provided prompts
Control Groups
Protocol-based standard therapy group: relaxed protocolised care (based upon published expert opinions) for 6 hours
Usual care group: no extra staffing, with all care as directed by bedside physician for 6 hours.
Research assistant collected data but provided no prompting.

14. ProCESS: Outcomes 60 Day Mortality: EGDT 21% Protocol Std 18.2%
Usual Care 18.9%
(P values )

15. ProCess Strengths Well designed
Attempted to tease out impact of “early” care vs “protocolised” care with 2 controls
Methods and statistical analysis defined and published a priori.
Recruited adequate numbers as planned for 80% power to detect 6–7% mortality reduction with p<0.05

16. ProCESS Weaknesses Inclusion criteria altered during trial:
 reduced amount of fluid required before "refractory hypotension" to 1L , but average vol still within Rivers‘ original definition of 20–30 ml/kg.
Interim recruitment target reduction as became clear much lower mortality (20% vs 30-45%)
Adherence to protocol was 88.1% in EGDT group and 95.6% in protocol-based standard therapy group.
Reflects “real life” but may reduce between group differences
Rivers 99% - ?assessment method

17. ProCESS Authors' Summary:
Adults with sepsis in the ED have ~20% 60-day mortality
Differing mortality rate found may reflect:
much improved awareness of sepsis care since Rivers’ paper
Rivers’ population older, more cardiac/liver disease, higher lactate
?more severe shock in Rivers’ group (but analysis of sickest in ProCESS still no difference)
No significant advantage (morbidity/mortality) of protocol- based resuscitation over bedside care provided by a dedicated team according to the treating physician's judgement.

18. ARISE: In a nutshell: 8y from conception to publication
Follow up to ProCESS (2014) and EGDT (2001)
EGDT protocol vs “usual care” in septic shock
One of 3 trials planned in “harmonized collaboration”
ProCESS (USA)
ARISE (Aus/NZ)
ProMISE (UK)

19. ARISE: In a nutshell: 51 centres Australia NZ ED presentations sepsis
EGDT vs usual care
90 day follow up for all-cause mortality 99%
= NO SIG DIFFERENCE
Only differences – they left ED earlier (for EGDT) and more vasopressor use (dictated by protocol)

20. Clinical question:
“In adult patients presenting to the emergency department with severe sepsis, does early goal-directed therapy (EGDT) reduce mortality at 90 days when compared with ‘usual care’?”

21. Study design: Multi-centre, unblinded, randomised controlled trial
Permuted block randomisation stratified by site to allocate eligible patients 1:1
Intention to treat analysis
Sample size calculation based on assumed in-hospital rate of death in the usual-care group of 28% with an increase to 38% by 90 days based on recent high quality studies patients provided a power of 85-90% to detect an absolute risk reduction of 7.6%

22. Study setting:
51 hospitals (45 in Australia or New Zealand and 6 centres in Finland, Hong Kong and the Republic of Ireland)
Non tertiary and tertiary centres
October April 2014
None of centres followed bundle based protocols/ScvO2 guided Rx

23. Study population
1600 patients enrolled EGDT: N=793 vs Usual care N=798 Similar baseline characteristics

25. Antibiotics at baseline: no difference
EGDT
Usual care
Antibiotics: time to admin (mins)
Median (IQR)
70 (38-114)
67 (39-110)
Appropriate antibiotics time to admin (mins)
91 (48-186)
89 (47-170)

26. Inclusion criteria: Suspected or confirmed infection AND
2 or more SIRS criteria AND
Either refractory hypotension OR hypoperfusion:
Refractory hypotension: SBP < 90 mmHg or MAP < 65 mmHg after 1 litre IV fluid bolus over 60 mins
Hypoperfusion: Lactate ≥ 4.0 mmol/L
AND first dose of IV antibiotics given prior to randomisation

27. Inclusion criteria:
First dose of IV antimicrobial Rx commenced prior to randomisation
Standard ED care
Lactate

28. Exclusion criteria:
Contra-indication to CVC insertion / blood products
Inability to start EGDT protocol within 1h of randomisation
Inability to complete 6h EGDT
Haemodynamic instability due to active bleeding
Pregnancy (confirmed or suspected)
In-patient transfer from other facility
Death deemed imminent and inevitable
Life expectancy <90 days due to underlying disease
Limitation of therapy order

29. ARISE: EGDT protocol Within 1 hour from randomisation
Adherence to protocols and resuscitationgoals = 95-99%

30. Control group: usual care
Usual resuscitation care:
Arterial line/CVL if clinically appropriate
ScVO2 measurement not permitted during 6 hour intervention period
Decisions about:
location of care delivery,
Investigations / monitoring / treatment
= at the discretion of the treating clinician

31. Comparison of interventions: Fluid

32. Comparison of interventions: Tranfusions

33. Comparison of interventions: Monitoring

34. Comparison of interventions: Vasopressor use
(Dobutamine use in 15.4% of patients vs. 2.6% in the usual care arm)

35. Primary outcome: no difference
Measure
EGDT
Usual Care AD
95% CI
P Value
90-day mortality Mean (SD)
18.6%
18.8%
-0.3%
-4.1 to 3.6%
0.90
AD = absolute difference; CI = confidence intervals; p = p value

36. Secondary outcomes: 2 differences
Measure
EGDT
Usual Care RR
95% CI p
Vasopressor support
76.3%
65.8%
1.16
1.09 to 1.24
< 0.001
ED LOS hours
1.4
2.0
n/a
ICU LOS days 
2.8
0.81
Hospital LOS days
8.2
8.5
0.89
Renal replacement Rx
13.4
13.5
0.99
0.77 to 1.27
0.94
Mechanical vent 30
31.5
0.95
0.82 to 1.11
0.52
28-day mortality
14.8%
15.9%
0.93
0.73 to 1.17
0.53
LOS = Length of Stay; ED = Emergency Department; ICU = Intensive Care Unit; RR = Relative Risk

37. Subgroup analyses: no outcome differences

38. Authors’ conclusions:
In critically ill patients presenting to the emergency department with early septic shock, EGDT did not reduce all-cause mortality at 90 days. 

39. Study strengths: Clinical relevance, high impact
Large multi-centre study, international, external validity
Difficult setting – crowded EDs with acute care pressures
Rivers’ original EGDT algorithm used - comparison
Pragmatic, reflects “real life” practice
clinician discretion in 'usual care' Mx/location care
Inclusion of centres with resource limitations
Timing of first dose antibiotics recorded (sensible enrolment criterion to avoid confounding effect of late administration)
Subgroup analyses reassuring regarding variation in mortality between countries
Very little loss to follow up – 99%
“Data massage” avoided by publication of statistical analysis plan before recruitment completed

40. Study weaknesses:
Lower APACHE scores than ProCESS and Rivers – subgroup analysis somewhat reassuring but small numbers small (150)
?Correct population recruited – ~70% in both groups with shock: - lactate levels similar.. Rivers group criticize much lower fluid volume given ?less sick population
Some elements of EGDT utilized in usual care group BUT there were significant differences in key elements of EGDT (PRC/dobutamine etc)
Real life mortality rates likely to be higher – study meant special team and early senior involvement for all
Low recruitment rate at all centres – 2 patients per month at largest (Austin)-reflects difficulty of research- but multicentre trials needed to ensure external validity, numbers and power to detect differences

41. ARISE: comparison with ProCESS:
Similar to ProCESS:
EGDT vs usual care
6h Rx regimen
Inclusion criteria identical
Differences from ProCESS:
Enrolled earlier (6h vs 12h)
Simpler – 2 arms
Aus/NZ/other countries
Non tertiary centres included

42. Comparisons of Rivers with ARISE/ProCESS:
Sicker patients:
= Baseline APACHE II – 20 in Rivers trial with higher rates of cardiac/liver
(ARISE = 15, ProCESS=20)
All 6 hour care given in ED in Rivers trial
Adherence to protocols 99%
After 6h, clinicians blinded to patient allocation
Mortality rates from sepsis MUCH higher:
= 30-46% vs around 20% for ProCESS
ARISE – 70% hypotense but received substantially less IV fluid than Rivers pts

43. Discussion points: 1 “Sicker patients” in Rivers
Attitudes to end of life care ANZ now
Dying not recruited
2 Hawthorne effect
Those being observed in a study try harder
3 Team focused on enrolled patient/ senior involvement early
No substitute for an excellent clinician standing there, joining the dots
4 Surviving sepsis campaign response to ProCESS and ARISE
– no evidence of HARM shown with invasive monitoring so: wait until all 3 studies results/meta-analysis is available
5 Ways to improve?
-QAS lactate/antibiotics – modified SIRS criteria – no WCC

44. Summary of impact of ARISE
1 Reinforces EARLY as the factor having impact
- supports ProCESS conclusion
- early adequate fluids and antibiotics and appropriate environment is key
2 Results achievable:
- across the world
- in non-tertiary settings
- in ED or ICU, without elaborate monitoring devices
3 Highlights work still to be done:
- ~ 20% mortality still
- significant delays to antibiotics

45. Sepsis fundamentals Early recognition
Early appropriate antimicrobial Rx
Early source control
Be careful at intubation
Right amount of fluid resuscitation
Early vasopressors
Early admission to monitored area and REASSESS

46. Or, in essence….

47. THANK YOU