SPLIT Header Image

Effect of a buffered crystalloid solution vs saline on acute kidney injury among patients in the intensive care unit: the SPLIT randomised clinical trial

Young P. JAMA. Published online October 07, 2015. doi:10.1001/jama.2015.12334

Clinical Question

  • In critically ill patients, does the use of a balanced crystalloid solution compared to normal saline effect the incidence of acute kidney injury?


  • Randomised controlled trial
  • Computer generated randomisation
  • Double blinded
  • Cluster, double cross over
  • Intention to treat analysis
  • No established statistical methodologies for prospectively determining sample sizes for this type of study and hence no sample size calculations were performed


  • 4 Intensive Care Units in New Zealand
    • tertiary ICU’s: 3 general and 1 predominantly cardiothoracic/vascular surgery
  • April 2014 – October 2014


  • Inclusion: all ICU patients requiring crystalloid fluid therapy as clinically indicated
  • Exclusion: end-stage renal failure requiring renal replacement therapy (RRT), currently requiring RRT or expected to require RRT within 6 hours; admission for consideration of organ donation; admission for palliative care; previous enrolment in trial
  • 2278 patient enrolled; 1162 into intervention arm, 1116 into control arm
  • Baseline characteristics were similar in both groups, Including: age; co-morbidities; time from ICU admission to first fluid; volume of fluid administered prior to enrolment and APACHE II score (14.1 in both groups)
    • High proportion of post elective surgery: 56% vs. 58%
    • High proportion of cardiac surgery (nearly 50% of study group)


  • Plasma-Lyte 148 (buffered crystalloid)


  • 0.9% sodium chloride

In both groups

  • Rate and frequency of fluid administration was determined by the treating clinician
  • Crystalloid treatment during intervention and procedures performed outside the ICU was the assigned study fluid as far as possible
  • No restrictions placed on other fluids or therapies
  • Buffered crystalloid group and saline group received similar volumes of study fluid (median [IQR]), 2000ml [1000-3500ml] vs 2000ml [1000-3250ml], P=0.63


  • Primary outcome: proportion of patients with AKI based on RIFLE criteria (‘injury’ or greater and based solely on creatinine component) within 90 days of enrolment: no statistical difference
      • 9.6% in intervention group vs 9.2% in control group; P=0.77
  • Secondary outcome: comparing patients who received Plasma-Lyte 148 with patients who received 0.9% sodium chloride – no statistical difference in any of the following
    • Acute Kidney Injury: Change in creatinine and incidence of AKI
      • Difference in serum creatinine measured before enrolment and peak creatinine in ICU: 0.21mg/dl vs 0.18mg/dl
      • Cumulative incidence of AKI (using both RIFLE and KDIGO definitions)
    • RRT: Requirement
      • In ICU: 3.3% vs 3.4%
      • No difference in indications for RRT initiation
      • Post hospital discharge: none
    • Mechanical ventilation: Proportion and duration of patients requiring mechanical ventilation:
      • 68.6% vs 67.7%
      • 15.3 hours vs 14.2 hours
    • ICU readmission during their index stay: 6.9% vs 5.1%
    • Length of stay: ICU stay: 1.5 days vs 1.47 days; Hospital stay: 7.45 days vs 7.33
    • All-cause mortality and cause-specific mortality: ICU: 6.6% vs 7.2%; Hospital: 7.6% vs 8.6%
  • Subgroup analysis

Authors’ Conclusions

  • 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.


  • Addresses a highly relevant question in practice and builds upon the knowledge from previous fluid trials
  • Intervention and control fluids were packaged in macroscopically indistinguishable 1000mls bag labeled “Fluid A” or “Fluid B”
  • Trial recruited and conducted on time and within 6 months
  • Statistical analysis plan published before completion of recruitment; including pre-defined subgroup analysis
  • 99.3% of eligible patients were enrolled into trial


  • Feasibility study; sample size and power calculations not possible
    • the data were consistent with a treatment effect that lies between a relative decrease of 33% and a relative increase of 17% in in-hospital mortality arising from the use of a buffered crystalloid instead of saline
  • Primary outcome data not available for 7.5%
  • Apart from cardiac surgery other Sub-groups all had small number of patients included
  • Main reason patients screened were not enrolled into study was that they did not require fluids and the amount of fluid administered within trial was small (median 2l) which raises questions about generalisability
  • More than 90% of patients were exposed to intravenous fluids before enrolment and the majority of pre-enrolment fluid was buffered crystalloid
  • Potential of ascertainment bias given that use of 0.9% sodium chloride is associated with hyperchloraemic acidosis and two-thirds of clinician were able to correctly identify fluid type
  • Physiological and biochemical separation were not reported (although manuscript being prepared – personal communication with author)

The Bottom Line

  • There was no difference in the incidence of acute kidney injury in patients receiving Plasma-Lyte 148 compared to 0.9% saline. Whilst more data from the trial is awaited, I will continue to use balanced solutions.
  • The study does provide reassurance that in elective or moderately sick intensive care patients, giving up to 2L 0.9% sodium chloride results in no increased risk of kidney injury compared with Plasma-Lyte 148

External Links


Summary author: @avkwong
Summary date: 9th October 2015
Peer-review editor: @stevemathieu75
Conflicts of Interest: None


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.