Get PulmCCM’s Weekly Email Update
Stay up-to-date in pulmonary and critical care. No spam.
“The world wavered and quivered and threatened to burst into flames.”
Intravenous fluid therapy is one of the most commonly-performed interventions in all of critical care medicine. Numerous trials over the last 20 years have attempted to identify the ideal fluid for those in the intensive care unit, with conflicting results. As the authors note, there is some consensus with regards to 4% albumin in traumatic brain injury and hydroxyethyl starch in kidney injury; that is, avoid these combinations! However, only in the last 5-7 years has there been more attention paid to crystalloid therapy; in particular the difference between 0.9% saline and ‘balanced’ solutions.
As touched upon previously, some of this interest has moved in-step with greater appreciation of Stewart’s Strong Ion Difference [SID] as a means to explain plasma acid-base chemistry. Given that balanced solutions replace strong, chloride anions with weaker conjugates, balanced solutions do not shrink the SID; consequently, with a larger SID, fewer protons emerge. In other words, there is less hyperchloremic, non-gap metabolic acidosis.
What they did
Adult patients admitted to one of 53 intensive care units in New Zealand or Australia between September 2017 and December 2020 were eligible. Further, the treating clinician must have deemed either normal saline or balanced crystalloid as appropriate resuscitation options; the patient’s expected-admission to the ICU must have been at least 3 days.
Patients were thus excluded if they had indications for a specific type of IV fluid resuscitation or if they had received more than 500 mL of IV fluid in the ICU. Other exclusions were imminent risk of death, life-expectancy less than 90 days, traumatic brain injury, risk of cerebral edema.
The patients were randomized and the trial fluids were double-blinded. Once randomized, the patient received only their assigned fluid for resuscitation for up to 90 days, but only within the ICU; the amount and rate of infusion was decided by the treating clinician.
Death, from any cause, within 90 days of randomization was the primary outcome. There were multiple secondary outcome measures that related to acute kidney injury, hemodynamic stability and respiratory failure. Further, death from any cause within the ICU, the hospital or within 28 days of randomization were also reported secondary outcomes. The analysis was performed as an intention-to-treat.
A sample size of 8800 was calculated to detect a 2.9% absolute risk reduction in mortality within 90 days, however, because of the coronavirus pandemic, the study was terminated early without knowledge of the outcome data.
What they found
5037 patients were randomized – 2522 to saline and 2515 to balanced crystalloid; approximately 96% of the randomized patients had data available for analysis. 45% of patients were admitted to the ICU directly from operating or recovery rooms, 79% were mechanically-ventilated and 42% had sepsis.
In the 24 hours prior to randomization, 55.5% of patients randomized to balanced fluids received at least 500 mL of saline, while only 23% of those randomized to saline received at least 500 mL of balanced fluids. As well, after randomization, 63% of patients randomized to balanced fluids received 500 mL or more of open-label saline, while only 3.5% of patients randomized to saline received at least 500 mL of open-label balanced fluids.
In the first week, those randomized to balanced fluids had a statistically-significantly higher arterial blood pH [i.e., by no more than a pH of 0.03] and lower serum chloride concentration [i.e., by 2-3 mmol/L]. Notably, serum potassium was essentially the same in both groups across the first week.
There was no statistically or clinically-significant difference in any of the primary or secondary outcomes.
With each additional randomized trial comparing balanced to 0.9% saline crystalloid resuscitation, the inter-trial evaluations become murkier, though seemingly less tendentious. In totality, the baseline patient characteristics, interventions and degree of fluid ‘cross-contamination’ are most similar between the two most recent trials; that is, BaSICS and PLUS.
Together BaSICS and PLUS randomized a little over 16,000 critically-ill patients to Plama-lyte 148 or 0.9% saline. Patients in the PLUS trial appeared sicker at baseline than BaSICS and stood to profit the most from the putative, beneficial effects of balanced resuscitation. The median APACHE II score was 19 in PLUS and 12 in BaSICS. Additionally, at baseline a larger fraction of patients were mechanically-ventilated and the mean creatinine was higher in PLUS [i.e., 1.44 mg/dL] as compared to BaSICS [i.e., 1.2 mg/dL].
The total fluid burden and ‘cross-contamination’ for each group is challenging to compare between the trials, but it appears that patients in PLUS received the most crystalloid in the first 7 days. Those randomized to 0.9% saline in PLUS garnered about 4.85 L of saline while those in the balanced group were infused with approximately 5.1 L of balanced solution [these amounts were closer to 3.0 L for both groups in BaSICS and 2.0 L in SMART]. Of note, however, a significant fraction of patients in both groups received additional, open-label, saline in the first 7 days [roughly 1.6 L to the saline group and 1.7 L to the balanced group], whereas there was very little open-label balanced solution provided to either group. In the 24 hours prior to randomization in PLUS, both groups, in totality, received roughly 1.0 L of saline and about 600 mL of balanced solution.
To approximate average fluid ‘cross-contamination,’ a patient in PLUS randomized to receive balanced solution was infused approximately 8.5 L of total crystalloid in the 24 hours prior to randomization plus the first 7 days of study; of this total amount, roughly 2.75 L [i.e., ~33%] was 0.9% saline. On the other hand, those randomized to 0.9% saline received roughly 8.1 L of total fluid in the same time-frame and only about 700 mL [i.e., ~ 9%] was balanced.
Given the, apparently, large chloride burden engendered via cross-contamination in PLUS as well as the early termination of the trial, proponents of balanced solutions may criticize the results. Yet, even so, the outcomes in PLUS echo those of BaSICS – which had less fluid cross-over – and in all three studies [SMART, BaSICS, PLUS] the average difference in chloride levels between the groups was on the order of 1-3 mEq/L in the first week. While SMART was able to achieve the least amount of fluid cross-contamination and also randomized nearly as many patients as PLUS and BaSICS combined, we recall that the primary end-point in SMART was composite – with none of mortality or worsening renal failure achieving significance in isolation. Therefore, the zeal for balanced fluids will lessen in the ICU, especially in those patients deemed to have no clear indication for either.
Lastly, there was no difference in potassium levels between the two groups in PLUS and, in fact, on day 2 serum potassium was significantly lower in those randomized to Plasma-lyte; therefore, holding balanced solutions for fear of raising potassium may be misplaced in the majority of critically-ill patients.