SPLIT: Saline vs. Plasma-Lyte in the ICU (2015)

“In this randomized clinical trial of critically ill adults, we did not find a significant difference in the incidence of acute kidney injury between patients receiving a buffered crystalloid and those receiving saline.”

— The SPLIT Investigators

1. Publication Details

  • Trial Title: Association Between a Buffered Crystalloid Solution vs Saline Solution and Acute Kidney Injury Among Critically Ill Patients: The SPLIT Randomized Clinical Trial.
  • Citation: Young P, Bailey M, Beasley R, et al; for the SPLIT Investigators and the ANZICS CTG. Association Between a Buffered Crystalloid Solution vs Saline Solution and Acute Kidney Injury Among Critically Ill Patients: The SPLIT Randomized Clinical Trial. JAMA. 2015;314(16):1701-1710. doi:10.1001/jama.2015.12334.
  • Published: October 27, 2015, in The Journal of the American Medical Association (JAMA).
  • Author: Paul Young, Ph.D.
  • Funding: Health Research Council of New Zealand.

2. Keywords

Intravenous Fluids, Crystalloid Solutions, Balanced Crystalloids, Saline, Plasma-Lyte, Critical Illness, Acute Kidney Injury (AKI).

3. The Clinical Question

In critically ill adults requiring crystalloid fluid therapy in the ICU (Population), does the use of a buffered crystalloid (Plasma-Lyte 148) (Intervention) compared to 0.9% saline (Comparison) reduce the incidence of acute kidney injury (AKI) (Outcome)?

4. Background and Rationale

  • Existing Knowledge: Saline (0.9% sodium chloride) was a very common intravenous fluid, but its high chloride content was known to cause hyperchloremic metabolic acidosis and was associated with renal vasoconstriction and acute kidney injury in animal and observational studies.
  • Knowledge Gap: Prior to this study, there were no large, multicenter, randomized controlled trials to definitively determine if using a buffered, lower-chloride crystalloid fluid improved kidney outcomes compared to saline in a broad population of critically ill patients.
  • Proposed Hypothesis: The authors hypothesized that the use of a buffered crystalloid solution (Plasma-Lyte) would result in a lower incidence of AKI compared to the use of saline.

5. Study Design and Methods

  • Design: A prospective, multicenter, double-blind, cluster-randomized, double-crossover trial.
  • Setting: Four intensive care units (ICUs) in New Zealand.
  • Trial Period: Enrollment from April 2014 to October 2014.
  • Population:
    • Inclusion Criteria: All adult patients (≥18 years) admitted to a participating ICU who were prescribed crystalloid fluid therapy.
    • Exclusion Criteria: Patients expected to be discharged within 24 hours, those with end-stage kidney disease on dialysis, or those admitted for kidney transplantation.
  • Intervention: Plasma-Lyte 148 (a buffered crystalloid) as the sole crystalloid fluid for all patients in the ICU.
  • Control: 0.9% saline as the sole crystalloid fluid for all patients in the ICU.
  • Management Common to Both Groups: The assigned study fluid was used for all crystalloid therapy (resuscitation, maintenance, and drug dilution). The ICUs were cluster-randomized to a fluid, and then crossed over to the other fluid after a 7-week period. This was repeated once.
  • Power and Sample Size: The trial was powered to detect a 2.2% absolute difference in the incidence of AKI, requiring 2252 patients.
  • Outcomes:
    • Primary Outcome: Proportion of patients who developed acute kidney injury (AKI), defined as stage 2 or 3 of the KDIGO criteria, within 90 days of enrollment.
    • Secondary Outcomes: Included the incidence of renal replacement therapy (RRT), peak creatinine level, and in-hospital mortality.

6. Key Results

  • Enrollment and Baseline: 2278 patients were included in the analysis (1152 assigned to buffered crystalloid, 1126 to saline). The groups were well-matched at baseline.
  • Trial Status: The trial was completed as planned.
  • Primary Outcome: There was no significant difference in the incidence of AKI between the buffered crystalloid and saline groups (9.6% vs 9.2%; P=0.77).
  • Secondary Outcomes: There were no significant differences in the incidence of RRT (3.3% vs 3.4%) or in-hospital mortality (7.6% vs 8.6%).
  • Adverse Events: The incidence of adverse events was similar in the two groups.

7. Medical Statistics

  • Analysis Principle: An intention-to-treat analysis was performed, accounting for the cluster-crossover design.
  • Statistical Tests Used: The primary outcome was analyzed using a mixed-effects logistic regression model.
  • Primary Outcome Analysis: The proportion of patients developing AKI was compared between the two groups.
  • Key Statistic(s) Reported: Adjusted Odds Ratio (aOR) for AKI with buffered crystalloid: 1.04 (95% CI, 0.80 to 1.36; P=0.77).
  • Interpretation of Key Statistic(s):
    • Odds Ratio (OR): An aOR of 1.04 suggests a 4% higher odds of AKI with buffered crystalloid, but this was not statistically significant.
    • Confidence Interval (CI): The 95% CI widely crosses 1.0 (from a 20% reduction to a 36% increase in odds), indicating no statistically significant difference.
    • P-value: The p-value of 0.77 is far above the 0.05 threshold, confirming the lack of a statistically significant difference.
  • Clinical Impact Measures: Not applicable as no benefit was shown for the primary outcome.
  • Subgroup Analyses: No significant differences were found in any of the pre-specified subgroups.

8. Strengths of the Study

  • Study Design and Conduct: The double-blind, cluster-randomized, double-crossover design was methodologically rigorous and minimized bias.
  • Generalizability: The pragmatic inclusion of nearly all ICU patients increases the external validity of the findings to a general ICU population.
  • Blinding: This was one of the few large fluid trials to successfully maintain blinding of clinicians to the treatment allocation.

9. Limitations and Weaknesses

  • Internal Validity (Bias): No major limitations to internal validity.
  • External Validity (Generalizability): The study was conducted in a specific healthcare system (New Zealand) which may differ from others. The patient population had a relatively low severity of illness and low mortality rate.
  • Other: The volume of study fluid administered was relatively small (median of <2 liters), which may have been insufficient to cause or prevent a significant amount of AKI. The study may have been underpowered to detect a smaller, but still potentially important, difference.

10. Conclusion of the Authors

“Among critically ill adults, the use of a buffered crystalloid compared with saline did not reduce the risk of acute kidney injury.”

11. To Summarize

  • Impact on Current Practice: The SPLIT trial was one of the first large, blinded RCTs to challenge the long-held belief that saline was harmful to the kidneys. Its neutral finding was surprising to many and generated significant debate. It suggested that for many general ICU patients receiving modest volumes of fluid, the choice between saline and a buffered crystalloid may not be critically important for renal outcomes.
  • Specific Recommendations:
    • Patient Selection: For a general population of critically ill adults requiring modest volumes of crystalloid fluid.
    • Actionable Intervention: This trial does not support the preferential use of Plasma-Lyte over saline for the purpose of preventing AKI.
    • Expected Benefit: No difference in the rate of AKI was demonstrated.
  • What This Trial Does NOT Mean: This trial does not mean that saline is harmless in all situations. The findings may not apply to patients receiving very large volumes of fluid (e.g., in major trauma or severe septic shock) where the effects of hyperchloremia could be more pronounced.
  • Implementation Caveats: The results provide reassurance that the use of saline for routine fluid therapy in a general ICU population is not associated with a high risk of AKI.

12. Context and Related Studies

  • Building on Previous Evidence: This trial provided high-quality RCT evidence that contrasted with the signals of harm seen in prior observational studies.
  • Influence on Subsequent Research: The neutral findings of SPLIT were a major impetus for the much larger SMART (2018) and SALT-ED (2018) trials. These later trials, using a different pragmatic design, did find a benefit for balanced crystalloids on a composite outcome (MAKE-30), suggesting that the patient population and the specific outcome measured are critical.

13. Unresolved Questions & Future Directions

  • Unresolved Questions: Does the choice of crystalloid matter in patients receiving large-volume resuscitation? Are specific subgroups of patients (e.g., those with sepsis or DKA) more susceptible to the effects of high-chloride fluids?
  • Future Directions: Research has continued with larger trials (SMART, BaSICS, PLUS) to clarify the role of different crystalloids in specific, high-risk populations and to understand the seemingly conflicting results between studies.

14. External Links

15. Framework for Critical Appraisal

  • Clinical Question: The question was highly relevant, addressing a fundamental and ubiquitous intervention in critical care.
  • Methods: The double-blind, cluster-randomized, crossover design was methodologically excellent and is a major strength of the trial.
  • Results: The trial had a clear and robustly negative result for its primary outcome. The lack of difference in any secondary outcome further strengthened this conclusion.
  • Conclusions and Applicability: The authors’ conclusion is a direct and fair interpretation of the data. The results are applicable to general ICU patients with moderate severity of illness receiving modest fluid volumes. The trial was a crucial piece of the puzzle that tempered enthusiasm for abandoning saline and highlighted the need for larger, more definitive trials in higher-risk populations.

16. Disclaimer and Contact

This summary is provided by the Academic Committee of ESBICM (ACE) to facilitate the understanding of this study; readers are advised to refer to the original trial document for a deeper understanding. If you find any information incorrect, or missing, or it needs an update or have a request for a specific critical care trial summary, kindly write to us at academics[at]esbicm.org.

Scroll to Top