Fluids

Bottom line: This abridged version of the full Cochrane review, which looked at 18 RCTs, demonstrates that there is low GRADE evidence for no statistically significant difference in mortality when using buffered solutions (i.e. Lactated Ringers) compared to non-buffered solutions (i.e. 0.9% NaCl) perioperatively. There was statistically significant differences in various metabolic variables: (a) Normal saline use resulted in a lower pH (b) Normal saline use resulted in a higher chloride concentration (c) Buffered solution groups had less negative base excess (d) serum bicarbonate was lower in non-buffered groups. There was limited evidence on the effects of fluid therapies on reducing post-operative organ failure. 

​

Major points:

​

1. Three trials across 267 participants showed 2.9% mortality in non-buffered solution groups and 1.5% mortality in buffered solutions. The results were not statistically significant and the sample size was small. The confidence in the effect estimate is low. 

​

2. There was no statistically significant differences in urine output, post-op serum Cr (other than a single trial), PaCO2, post-operative nausea and vomiting, serum sodium, potassium, glucose, lactate, measures of coagulation, LOS, blood loss or transfusion requirements. Some metabolic derangements were statistically significant (pH, bicarb, chloride, base excess) but the clinical significance of these are not known. 

​

3. Given the millions of patients that receive perioperative fluids each year, there is a surprising paucity in high-quality trials to guide their use. The literature that is available is across a variety of patient populations undergoing a multitude of surgeries. The timing and volume of fluids administered were not reported. Fluid administration should be individualized to each patient as it is difficult to prescribe a "one-size fits all" approach especially given the lack of data. 

​

4. A well-designed and adequately powered trial is needed to be able to detect any clinical differences. It should measure: mortality, recovery, lOS, organ dysfunction and quality of life. The SOLAR trial is one such trial looking at NaCl vs Ringer's Lactate in post-operative complications - it is scheduled to be completed in 2022. 

​

Selection of Intravenous Fluids

Kwong YD, Liu KD. Selection of Intravenous Fluids. Am J Kidney Dis. 2018;72(6):900-902. doi:10.1053/j.ajkd.2018.05.007

Bottom line: This 3-page nephrologist-centered review article discussed the findings and implications of the Vanderbilty University-based SMART and SALT-ED trials which compared balanced crystalloids (Ringer's Lactate) compared to 0.9% NaCl in critically ill and non-critically ill patients, respectively.

SMART reduced 30-day mortality from 29.4% to 25.2% (especially in septic patients) with the use of balanced crystalloids while SALT-ED reduced the incidence of persistently decreased kidney function upon discharge with the use of balanced solutions. Both studies showed decreased MAKE-30 (composite of death, new use of RRT and persistently decreased kidney function at discharge) by 1.1% in SMART and 0.9% in SALT-ED with the use of balanced solutions.

​

Given this emerging evidence, it is likely that nephrologists will trend towards favouring balanced crystalloids in fluid resuscitation - especially in the critically ill with sepsis. The potential benefit of balanced crystalloids must be weighed against the increase in cost ($4.50/L Ringer's/Plasmalyte vs $2 for normal saline) especially given the likely large NNT given how many patients (~15 000) were required to adequately power both studies to see an effect. 

​

​

Balanced Crystalloids versus Saline in Noncritically Ill Adults. (SALT-ED trial)

Self WH, Semler MW, Wanderer JP, et al. Balanced Crystalloids versus Saline in Noncritically Ill Adults. N Engl J Med. 2018;378(9):819-828. doi:10.1056/NEJMoa1711586

Bottom line: This single-centre, pragmatic, multiple-crossover trial of 13,347 patients receiving at least 500mL of either balanced crystalloids (Ringer's Lactate or Plasmalyte) or normal saline (before hospitalization outside of the ICU) showed no difference in the primary outcome: the number of hospital-free days by day 28 (median 25 days for both groups). The median amount of fluid administered was 1079mL. 

Major points:

​

1. The balanced crystalloid group had a 0.9% lower absolute risk of MAKE-30 (the secondary outcome, 4.7% vs 5.6%). MAKE-30 is a composite of in-hospital death, new RRT or persistent renal dysfunction (Cr >200% from baseline upon discharge. This corresponds to a NNT of 111. The patients who presented to the ED with renal dysfunction or elevated chloride received the most benefit from balanced crystalloids.

​

2. The balanced crystalloid group had lower chloride and bicarb concentrations which persisted several days into hospitalization. The balanced crystalloid group also had a lower incidence of acidemia. 

​

3. The incidence of Stage 2 or greater AKI was 8.0% in the balanced group and 8.6% in the saline group - this was not statistically significant. Those who presented to the ED in AKI were more likely to see a resolution of their AKI during their hospital admission when given balanced crystalloids (28% vs. 37.6%). 

​

4. The authors suggest that despite the relatively large NNT (111 patients to reduce one major adverse kidney event at 30 days) given that millions of people receive NaCl each year, this finding can still have a profound impact. 

Balanced Crystalloids versus Saline in Critically Ill Adults (SMART trial)

Semler MW, Self WH, Wanderer JP, et al. Balanced Crystalloids versus Saline in Critically Ill Adults. N Engl J Med. 2018;378(9):829-839. doi:10.1056/NEJMoa1711584

Bottom line: This pragmatic, unblinded, cluster-randomized, multiple-crossover trial in 15802 ICU patients receiving either normal saline or balanced crystalloid (Lactated Ringer's) showed a statistically significant 1.1% reduction of MAKE-30 (major adverse kidney events at 30 days - the composite of death, new RRT or persistent renal dysfunction at 30 days) in patients receiving balanced solutions (15.4% in saline group vs 14.3% in balanced crystalloids group). This corresponds with a NNT of 94. 

​

Major points:

​

1. Patients receiving balanced crystalloids had less incidence of hyperchloremia (24.5% vs 35.6%) and plasma bicarb < 20 mmol/L (35.2% vs 42.1%). These differences were greater for those receiving larger volumes of fluids.

​

2. The primary outcome, risk of MAKE-30, was greater in patients with sepsis and those receiving larger volumes of fluids.

​

3. The balanced crystalloid group had a mortality rate of 10.3% compared to the saline group (11.1%) - this was not statistically significant. There were no differences in secondary clinical outcomes: ICU-free days, ventilator free days, vasopressor-free days, days alive & free of RRT. There was no differences in secondary renal outcomes: new RRT, persistent renal dysfunction or AKI > or equal to Stage 2. 

​

4. Despite an NNT of 94 to prevent one MAKE-30 in critically ill adults, over the 5 million yearly ICU admissions prescribing balanced crystalloids over normal saline could have a substantial effect. This effect would be even greater in septic patients or those patients requiring larger volumes of fluids. 

Current concepts of fluid management in enhanced recovery pathways.

Makaryus R, Miller TE, Gan TJ. Current concepts of fluid management in enhanced recovery pathways. Br J Anaesth. 2018;120(2):376-383. doi:10.1016/j.bja.2017.10.011

Bottom line: The goals of perioperative fluid therapy are to enhance patient, outcomes, decrease complications and decrease hospital LOS. The perioperative physician can intervene pre-operatively, intra-operatively and post-operatively to achieve these goals. 

​

Pre-operatively, PO hydration should be encouraged right up until 2 hours pre-operatively. Hydration in this period with complex carbohydrate drinks has multiple benefits (including reducing anxiety, PONV and LOS) while not increasing the risk of aspiration. 

​

Intra-operative fluid management balances the risk of too little fluids (i.e. hypotension and low perfusion states) with excess fluids (i.e. edema and fluid excess states). Ideally, there is a happy medium of euvolemia that is the goal. Although goal-directed therapy (GDT) has not shown any benefit in patient outcomes, it is not associated with harm, has the potential to improve outcomes by maintaining hemodynamic stability, and is currently embedded in most institutions' Enhanced Recovery Programs (ERP). More well-designed studies are needed on GDT. 

​

Post-operative fluid management focuses on encouraging early PO hydration (over IV fluids). While anuria is worrisome and should be treated accordingly, oliguria should be considered a normal response to the stress of surgery and should not be treated with aggressive IV fluids - some studies suggest a threshold of 0.3mL/kg/hr in major abdominal surgery.

​

​

Perioperative Fluid Therapy for Major Surgery

Miller TE, Myles PS. Perioperative Fluid Therapy for Major Surgery [published correction appears in Anesthesiology. 2020 Feb;132(2):405]. Anesthesiology. 2019;130(5):825-832. doi:10.1097/ALN.0000000000002603

Bottom line: Perioperative fluid therapy strives to maintain tissue and electrolyte homeostasis while avoiding salt and water excess. Historically, wide variations in clinical practice have made elucidation of the optimal fluid management strategy challenging with overall trends changing with the times. Emerging evidence is pointing to an individualized plan that considers patient risk and surgical risk. Patients with hypotension or evidence of hypovolemia should be fluid resuscitated until they are no longer fluid responsive before the implementation of vasopressors/inotropes.

Major points:

1. Preoperatively, patients should be encouraged PO intake up until 2hr pre-op with some institutions even allowing PO intake up until entry into the OR suite (although evidence is lacking for the latter). A complex carbohydrate drink pre-op has considerable benefits including improving patient satisfaction, thirst, hunger and PONV as well reducing post-operative hyperglycemia – as risk factor for post-op infections.

2. When assessing fluid responsiveness the goal is to identify whether additional IV fluids will increase cardiac output – taking advantage of the steep part of the Frank-Starling curve. Stroke volume (SV, as measured via pulse pressure variation or even echocardiography) is a better indicator of fluid status than blood pressure. Systematic review of 50 studies found physical exam was not predictive of fluid responsiveness, CVP < 8mmHg was predictive, as was vena cava diameter variation on U/S. The best predictor of fluid responsiveness was SV measurement during passive leg raising test – which transfers approximately 300mL into the RA and mimics a fluid bolus.

3. Intraoperative fluid management previously favored liberal approaches (>6L intra-op) and more recently much more restrictive approaches (e.g. zero-balance post-op). With emerging evidence of harm from zero-balance approaches (i.e. increase risk of AKI in RELIEF trial, using 1.7L vs 3L in liberal), the new recommendation is a moderately-liberal approach with 1-2L positive post-op (aka 3L crystalloid for a 3-4 hr major procedure, or 1-2L crystalloid for a lap chole).

4. Goal-directed therapy (GDT) harnesses the concept that the body strives to maintain pressure at the expense of flow – by monitoring SV as opposed to BP, hypovolemia can be recognized sooner. 20 years ago, many small studies on GDT showed benefit. More recent studies have failed to showed benefit – likely because modern GDT is embedded in ERAS pathways which have already done so much to improve outcomes compared to two decades ago. Fluid status should be optimized before a vasopressor/inotrope is added.

5. Post-operatively, early oral intake as per ERAS pathways is recommended. As long as PO intake is tolerated, IV fluids should be discontinued within 24 hrs – keeping in mind that in RELIEF, post-operative patients with restrictive fluids therapy (0.8 ml/kg/hr vs 1.5) had a near doubling of AKI).  

6. Isotonic, balanced crystalloids should be used whenever possible – due to the risk of hyperCl, metabolic acidosis and increased risk of kidney injury with 0.9% NaCl. The debate between colloid and crystalloid continues except in major hemorrhage, where blood products as per institutional massive transfusion protocols should be used.

Effects of Fluid Resuscitation With Colloids vs Crystalloids on Mortality in Critically Ill Patients Presenting With Hypovolemic Shock (CRISTAL trial)

Annane D, Siami S, Jaber S, et al. Effects of fluid resuscitation with colloids vs crystalloids on mortality in critically ill patients presenting with hypovolemic shock: the CRISTAL randomized trial [published correction appears in JAMA. 2013 Mar 12;311(10):1071. Régnier, Jean [corrected to Reignier, Jean]; Cle'h, Christophe [corrected to Clec'h, Christophe]]. JAMA. 2013;310(17):1809-1817. doi:10.1001/jama.2013.280502

Bottom line: This international, pragmatic, international trial of 2857 critically ill patients randomized them to either: any crystalloid (control group) or any colloid (experimental group) for resuscitation (if any of hypotension, low filling pressures or cardiac index or signs of tissue hypoperfusion were present). The study showed that the 28-day mortality – the primary outcome - did not differ between the two groups while 90-day mortality (secondary outcome) was lower in the colloid group (30.7% colloid vs 34.2% in crystalloid). The authors consider this finding exploratory only and recommend more research on efficacy. 

Major points:

1. The colloid group received significantly less fluids over the course of their ICU stay (2L vs 3L). There was however no difference between the two groups in mean BP, urine output, chest x-ray or weight during the 24 hrs. There was no difference in the total amount of blood transfused.​

2. The colloid group had significantly more days alive off vasopressor therapy within the first 7 days (5.0 vs 4.7 days) and 28 days (14.6 days vs 13.5 days). They also had significantly more days alive without mechanical ventilation in 7 days (2.1 days vs 1.8 days) and 28 days (14.6 days vs. 13.5 days). There were no differences in any of the other secondary outcomes including number of days alive without receiving RRT and days without organ system failure.

3. The majority of the crystalloid group (86%) received 0.9% NaCl. The majority of the colloid group (70%) received hydroxyethyl starches. There was no increased requirement of RRT in the colloid group – as was demonstrated in previous studies.

4. The survival curves between the two groups start to separate at 3 weeks – there is no clear explanation for this.

5. The findings of this study should be considered exploratory until further research is done on efficacy.

Assessment of hemodynamic efficacy and safety of 6% hydroxyethylstarch 130/0.4 vs. 0.9% NaCl fluid replacement in patients with severe sepsis: The CRYSTMAS study

Guidet B, Martinet O, Boulain T, et al. Assessment of hemodynamic efficacy and safety of 6% hydroxyethylstarch 130/0.4 vs. 0.9% NaCl fluid replacement in patients with severe sepsis: the CRYSTMAS study. Crit Care. 2012;16(3):R94. Published 2012 May 24. doi:10.1186/cc11358

Bottom line: This 2012 randomized, controlled, double-blind, prospective, multicenter study out of France and Germany compared a total of 196 patients admitted into the ICU with severe sepsis receiving either a colloid (6% HES 130/0.4) or crystalloid (0.9% NaCl). The primary outcome – volume need to achieve hemodynamic stability – was significantly lower in the HES group (1379mL vs. 1709mL). The two groups had comparable kidney function at study completion.

Major points:

1. There was no statistically significant difference in LOS in ICU, LOS in hospital, vitals signs, hemodynamic parameters, or time needed to achieve HDS (11.8 hr in HES group vs 14.3 hrs in NaCl group).

2. Site of sepsis, causative organism and age was associated with increased mortality. Choice of fluid was not. Mortality at day 28 was lower in NaCl group (25.3% vs 31%), however this was not statistically significant. Mortality at day 90 was also not statistically significant.

3. The two groups had comparable kidney function at study completion as per RIFLE and AKIN criteria (although patients with reduced baseline kidney function were excluded). There was no differences in mean serum creatinine, or urinary biomarkers for AKI. There were no difference in transfusion rates or in laboratory coagulation parameters. Both groups experienced similar rates of itching (3.0% vs 3.1%).

Hydroxyethyl Starch 130/0.42 versus Ringer's Acetate in Severe Sepsis

Perner A, Haase N, Guttormsen AB, et al. Hydroxyethyl starch 130/0.42 versus Ringer's acetate in severe sepsis [published correction appears in N Engl J Med. 2012 Aug 2;367(5):481]. N Engl J Med. 2012;367(2):124-134. doi:10.1056/NEJMoa1204242

Bottom line: This 2012 Scandinavian, investigator-initiated, multi-center, blinded, stratified, parallel-group clinical trial study compared 798 patients admitted with severe sepsis (or septic shock) to fluid resuscitation with either HES 130/0.4 or Ringer’s Acetate. The study showed that the HES group had an 8% higher absolute risk of mortality at 90 days (primary outcome, 51% vs 43%), were more likely to receive RRT (22% vs 16%), had fewer days without RRT and fewer days alive outside of hospital.

Major points:

1. The increased risk of 90-day mortality with HES resuscitation corresponds to a number needed to harm of 13. The authors postulate that HES has a high fraction that is not metabolized has been shown to deposit in the kidney, liver and bone marrow. This could explain why the survival curve separates only after ~20 days.

2. Despite previous trials showing that colloids were more effective intravascular volume expanders than crystalloids, both groups received median cumulative fluid of 3000mL.

3. Although the primary outcome was a composite outcome (death at 90-days or dialysis at 90 days), the effect was due to mortality at 90-days as only 1 patient required dialysis at 90 days.  

4. There were no differences in any of the secondary outcomes, including: death at 28 days, severe bleeding, severe allergic reactions, days alive without mechanical ventilation.

Passive leg raising for predicting fluid responsiveness: a systematic review and meta-analysis.

Monnet X, Marik P, Teboul JL. Passive leg raising for predicting fluid responsiveness: a systematic review and meta-analysis. Intensive Care Med. 2016;42(12):1935-1947. doi:10.1007/s00134-015-4134-1

Bottom line: This 2016 meta-analysis of 21 studies looked at the ability of the passive leg raise (PLR) test to predict fluid responsiveness as measured by cardiac output (CO) and arterial pulse pressure (PP) across a total of 991 patients and 995 fluid challenges in the ICU setting. The authors conclude that when combined with CO, PLR has excellent sensitivity and specificity for determining which patients would respond to a fluid challenge (Sens 81%, Spec 91%).  When combined with PP, the sensitivity and specificity was poor (56% sens, 83% spec).

Major points:

1. A PLR increases systemic vascular resistance, promotes venous return, thereby increasing right and left ventricular preload to a sufficient degree as to challenge the preload responsiveness of both ventricles. A threshold of 10% increase in CO induced by PLR is suggestive of fluid-responsiveness.  The reliability of the PLR is dependent on the accuracy of the method used to measure the CO.

2. PLR was highly predictive of fluid-responsiveness without taking into account the CVP. This is important as previous studies have suggested that PLR should increase CVP by 2mmHg in order to provide adequate preload.

3. Only the CO (or surrogate) should be used when assessing the PLR, not the pulse pressure. Furthermore, PLR should not lead to reflexive administration of fluid. Rather, it is a diagnostic tool to be used to assess which acute circulatory failure patients, as indicated by low arterial pressure or signs of poor perfusion, would see an increase in cardiac output with the administration of a fluid bolus – all without the need to actually administer any fluid. PLR is thus a “virtual fluid challenge.

4. PLR is a diagnostic tool. It has never been demonstrated to improve survival. Fluid responsiveness itself is not associated with improved outcome.