Fluid balance in ECMO patients


Schmidt M, Bailey M, Kelly J, Hodgson C, Cooper DJ, Scheinkestel C, et al. Intensive care medicine. 2014;40(9):1256-66.

Clinical question

Is early fluid balance status (within the first three days of ECMO initiation) associated with outcome in adult patients?


Retrospective observational study


Single- centre 45 bed ICU in Australia (Alfred Hospital, Melbourne)

March 2006- March 2013


Adult ICU patients supported by either VV or VA ECMO, for indications as determined by treating Intensivist or cardiac surgeon

Excluded 21 patients who didn’t survive to day 3, and 2 patients who had incomplete data from referral hospital.

172 patients analysed (VA: 115, VV: 57)

Subdivided into those who received renal replacement therapy (RRT) during their ICU stay (n= 103) and those who didn’t (n= 69)


Participants assigned to cohorts and analysed according to fluid balance.

Daily fluid balance defined as: negative if fluid lost> fluid administered, and positive if fluid lost<fluid administered, over 24hour period


Positive daily fluid balance at ECMO day 3

Primary outcome: mortality at 90 days after 3 days on ECMO

  • OR 4.02 (95% CI 1.49-10.82, p=0.006

Graph showing mean daily fluid balance during ECMO according to 90-day outcome (white square= 90-d survivors, black circle= non-survivors). Data expressed as mean+/-SD. * p &lt;0.01

Graph showing mean daily fluid balance during ECMO according to 90-day outcome (white square= 90-d survivors, black circle= non-survivors). Data expressed as mean+/-SD. * p <0.01

Secondary outcomes

Hospital mortality:

  • OR 3.74 (95% CI 1.44-9.70, p=0.007)

Time to death from ECMO day 3 up to day 90:

  • HR 4.30 (95% CI 1.80- 10.31, p=0.001)

Time to hospital death:

  • HR 4.31 (95% CI 1.79-10.36, p=0.001)

Post-hoc analysis: effect of AKI and CRRT use on outcome

90-day mortality AKI vs non-AKI:

  • 31% vs 15% (p<0.001)

90-day mortality CRRT during day 1-3 vs no CRRT day 1-3

  • 3.17 (95% CI 1.32-7.61, p=0.01)

Author’s conclusion

Early positive fluid balance, especially at day 3, is a robust independent predictor of 90-day mortality during ECMO, regardless of the primary diagnosis, AKI, or RRT use

ECMO patients requiring CRRT are at greater risk of 90-day mortality.


Answers a clinically relevant question.

Despite reasonable data in children, data on fluid balance in adult ECMO patients is limited. This is the largest study to date in the field, and opens up the need for a multicentre trial.

Outcome measures included medium-term survival (including 90 day and hospital survival), and adjusted for important variables such as bleeding events, AKI, CRRT use.

Outcome adjusted for confounding factors such as AKI or RRT use (which may affect outcome in their own right)

Analysis used propensity- scoring to assess positive fluid balance as binomial or continuous variable.


Retrospective observational design

Type of fluid administered was not recorded. Authors acknowledge this, but note that starches were not available during the study.

Cumulative fluid balance during ICU stay (rather than just mean daily balance) was not reported (?maybe just as relevant to outcome)

Low overall mortality rate (23% VV, 25% VA) may limit generalisability


Overall this is an excellent study looking at the (surprisingly) previously undefined issue of fluid balance in adult ECMO patients. Early positive fluid balance in adult ECMO patients is associated with worse outcome, which has implications on how aggressively we should be removing fluid in the first days of ECMO support.  A prospective, multicentre trial on the issue is certainly warranted.

Click here for abstract of the original article (not full text)

Summary author: Pete McCanny

Reviewers: Sean Scott, Steve Morgan