1. Journal Club: Initiation Strategies for Renal Replacement Therapy (RRT) in the ICU
Toby Chanin

2. Background
Acute Kidney Injury/failure is a common condition in Critical care
RRT has different modes and methods
Standard indications for RRT:
Symptomatic uraemia
Resistant fluid overload
Electrolyte abnormalities
Acid/base disturbance unresponsive to medical Mx
Drug filtering

3. Complications of RRT As with any intervention downsides:
Hypotension/haemodynamic instability
Clotting - increases blood loss
Bleeding risk
Hypothermia
Access issues
Therefore having clear indications for initiation is key to avoid harms.

5. Paper NEJM 07/2016 - impact factor - 59.558 (2015)
Aim to examine outcomes for patients with AKI #3 who receive early vs delayed RRT
Definition used of AKI #3 is that of Kidney Disease: Improving Global Outcomes classification (KDIGO)
[Cr]p x3 of baseline or>354µmol/L
Or UO <0.3ml/kg over 24 hr
Or anuria for 12 hours

6. Hypothesis
Delayed RRT in patients with AKI#3 will have a 15% lower mortality when compared with early RRT
This was based on their evaluation of current data
Expected mortality to be approximately 55%

7. Study design
AKIKI - artificial kidney initiation in kidney injury trial
Unblinded, prospective, multi-centre, open- label, two-group, randomised trial
Across 31 ICUs in France from 09/13-01/16
Written consent not required

8. Inclusion criteria Age>18 yrs
Had Dx of AKI 2o to ATN of ischaemic or toxic injury
Received Mechanical ventilation and/or Catecholamine infusion
Meet KDIGO criteria for AKI #3
Once identified they were randomised within 5 hours.

9. Exclusion Criteria
Excluded if any of the following were present at enrollment:
BUN> 40mmol/L
K+ >6 or >5.5 after medical mx
pH <7.15
Acute pulmonary oedema requiring O2 >5L/min to keep SpO2>95% or FiO2>50%

10. Intervention Early initiation of RRT
Had to be started ASAP after randomisation
Needed to be within 6 hours of identification of AKI #3
Decision regarding mode, timing and method of RRT not controlled between sites.

11. Comparator
Delayed initiation group were patients with AKI#3 who went on to develop:
Any of the previous exclusion criteria
BUN> 40mmol/L
K+ >6 or >5.5 after medical mx
pH <7.15
Acute pulmonary oedema requiring O2 >5L/min to keep SpO2>95% or FiO2>50%
Or remained oligouric/anuric for 72 hours.

12. Outcome Patients followed up for 60 days after randomisation
Primary outcome was survival at 60 days.
Secondary measures
Need for RRT in the delayed group
Number of RRT-free days
Number of dialysis cath free days
Number of MV-free days
Number of Vasopressor-free days
SOFA score at day 3 and 7
LOS in ITU and Hospital
Rate of nosocomial infection

13. Statistical Power
Study was powered to detect a 15% difference in mortality between the groups
Required 560 patients

14. Results 5528 eligible patients - 620 were randomised
80% of patients were septic
63% had been exposed to nephrotoxic agents

16. Primary Outcome 614 patients’ had data available at 60 days
303 had died - Overall mortality was 49.1%
No significant difference between groups
150 in early group
153 in delayed

17. Post-hoc analysis Further analysis of primary outcome
Broke down delayed group into those who needed RRT and those who did not
Revealed:
Never receiving RRT mortality of 37.1%
Early RRT mortality of 48.5%
Delayed with RRT mortality of 61.8% (p<0.001)

18. Secondary Outcomes No of delayed group needing RRT - 157 (51%)
Vs. 98% of early group (p<0.001)
Number of RRT free days
Delayed 19 vs early 17 (p<0.001)
No significant difference in requirement for respiratory or cardiovascular support
No significant difference in LOS
Significant difference in vasc. catheter related infections
30 (10%) in early vs 16 (5%) in delayed (p<0.03)

19. Discussion
Main aim of the study was really to try to use AKI #3 as early predictor as an advancement on traditional indications
No significant difference between the intervention groups in mortality
Perhaps instead using earlier targets on the traditional indicators may yield different results?

20. Discussion Delayed group had significantly fewer RRT sessions.
Has benefits of cost and time saving as well as obviating the complications risks.
Study not powered to breakdown delayed group - significant subgroup
Also how can you Identify which members of the 2nd group need earlier RRT?
Perhaps there is a way to identify trends of creatinine rise and identify earlier people who will need RRT

21. Limitations Investigators identify several:
Underpowered to identify a small mortality difference
Not using Kt/V to control RRT dosing
Only applicable to critically ill patients with AKI #3

22. Limitations Quite narrow inclusions and exclusion criteria
Unable to control for method of RRT used
Investigators recognised this and went on to state that 50% of pts received intermittent HD and only 30% had continuous HF alone
This may account for a higher mortality as HD is more haemodynamically destabilising vs HF

23. Interpretation
Investigators are cautious to ensure that the readers do not conclude, from this paper, that a ‘wait and see’ approach is safe.
My own interpretation would be that perhaps there is a potential to identify patients who may recover alone and those who we might be able to predict if they were worsening.

24. Conclusions
KDIGO AKI #3 is not a reliable single- predictor for the need to start RRT.
RRT therapy is not an intervention that should be started lightly given the high mortality attached to it.

25. Questions/comments