Fluid and Hyponatremia Management

By Craig B. Whitman, Pharm.D., BCPS, BCCCP

Reviewed by Serena A. Harris, Pharm.D., BCPS, BCCCP; Ohoud A. Aljuhani, Pharm.D., BCCCP; and Abigail D. Antigua, Pharm.D., BCCCP

LEARNING OBJECTIVES

1. Evaluate intravascular volume status and administer intravenous fluids (IVFs) to a critically ill patient.
2. Develop a plan to administer the appropriate IVF choice in a critically ill patient according to evidence from the literature.
3. Justify the use of peripheral administration of hypertonic saline and vasopressin antagonists to correct hyponatremia in
a critically ill patient.
4. Demonstrate the role of desmopressin in preventing sodium overcorrection when treating hyponatremia.

INTRAVENOUS FLUID THERAPY

ABBREVIATIONS IN THIS CHAPTER
Introduction
AKI Acute kidney injury
IVC Inferior vena cava Intravenous fluid (IVF) therapy has been used in clinical medicine
IVF Intravenous fluid since the 1830s during the European cholera epidemic (Severs 2015).
LOS Length of stay Many indications exist for IVF therapy in the critically ill population.
RRT Renal replacement therapy Intravascular volume replacement is the most common reason for
SIADH Syndrome of inappropriate using large volumes of IVF in the ICU, with sepsis, trauma, burn,
antidiuretic hormone and perioperative volume loss requiring replacement. Hypovolemia
TBI Traumatic brain injury occurs with reduction of extracellular volume and requires replace-
ment with IVF. Electrolyte repletion, management of acidosis, and
Table of other common abbreviations. prevention and management of drug-induced toxicities are other
examples of common indications for IVF (Severs 2015). Many IVFs
have been developed to optimize composition and administration in
patients requiring treatment. These solutions are classified as crys-
talloids and colloids. Crystalloid solutions contain small solutes with
various amounts of water. The prototypical crystalloid solution is
0.9% sodium chloride (normal saline), with lactated Ringer solution
and other various “balanced crystalloid” solutions available for IVF
therapy. Intravascular volume expansion from normal saline is main-
tained for up to 6 hours, with close to one-half of the infusate located
in the interstitium (Severs 2015). Colloid solutions contain large mole-
cules that more effectively maintain intravascular volume by exerting
an oncotic pressure. Available colloids include iso-oncotic (4%–5%)
and hyperoncotic (20%–25%) albumin solutions, semisynthetic col-
loids like hydroxyethyl starch, dextrans, gelatins, and blood products.
Despite the longstanding use of available IVFs, considerable
debate exists regarding the optimal fluid choice, dose, and method
of administration in the critically ill population. This chapter aims to
discuss recent advances in knowledge of IVF therapy and provide
insight into optimal use of various fluid choices in this population.

CCSAP 2018 Book 3 • Fluids and Nutrition/GI and Liver Disorders 7 Fluid and Hyponatremia Management
Assessment of Volume Status or large body habitus. In general, static measurements of
in the Critically Ill Patient intravascular blood volume (e.g., central venous or pulmo-
Assessing intravascular volume status and fluid responsive- nary artery occlusion pressure) should be avoided as the
ness in critically ill patients is one of the more challenging sole assessment method (Rhodes 2017; Kalantari 2013). To
aspects of managing IVF therapy in this population. Appropriate optimize therapy, intravascular volume should continually be
assessment is important because it may mitigate a positive assessed and reassessed in patients who receive IVF.
fluid balance, which may be associated with increased mor-
tality (Malbrain 2014). A variety of methods including physical IVF Administration in Critically Ill Patients
examination, chest radiography, laboratory values, and inva- Fluid may be administered in a variety of ways in critically ill
sive monitoring can be used to assess volume status. There patients. The 12th Acute Dialysis Quality Initiative Conference
is no universally accepted method to assess intravascular workgroup has proposed definitions pertinent to fluid adminis-
volume in critically ill patients. Physical examination findings tration (Hoste 2014). Fluid bolus is a rapid infusion to correct
should be used during the initial evaluation of intravascular intravascular volume status in shock states that consists of at
volume status and combined with dynamic measurements least 500 mL over a maximum of 15 minutes. A fluid challenge is
of blood volume. The Surviving Sepsis Campaign guide- a rapid infusion to correct hemodynamic instability, administer-
lines recommend using dynamic variables to assess fluid ing 100–200 mL over 5–10 minutes followed by reassessment
responsiveness, including passive leg raises, pulse pres- (Hoste 2014). The Surviving Sepsis Campaign recommends
sure variation, and stroke volume variation (Rhodes 2017). 30 mL/kg as the initial fluid challenge in patients with sep-
Esophageal Doppler monitoring and ultrasound measure- sis (Rhodes 2017). Fluid infusion is the continuous delivery of
ment of the inferior vena cava (IVC) can also be considered IVF to maintain homeostasis, replace losses, or prevent organ
if the equipment is available and the operator has expertise injury. Maintenance infusion is administration to provide flu-
in its use. Esophageal Doppler monitoring may be limited ids for patients who cannot maintain their needs by the oral or
because it requires the patient to be sedated (Kalantari 2013). enteral route. For individuals without ongoing losses, a maxi-
The accuracy of IVC ultrasonography may be affected by mum rate of 1–2 mL/kg/hour is recommended (Hoste 2014).
elevated intra-abdominal pressure (e.g., post-laparotomy) Four distinct phases of fluid therapy in resuscitation have been
proposed: rescue, optimization, stabilization, and de-escalation
(Rewa 2015). The rescue phase occurs with hemodynamic
BASELINE KNOWLEDGE STATEMENTS
instability and associated impaired organ perfusion, resulting
in life-threatening shock. Intravenous fluids are administered
Readers of this chapter are presumed to be familiar using boluses in the rescue phase. Optimization is when the
with the following: patient is no longer at imminent risk of life-threatening shock
• Understanding of intravascular fluid composition, but requires fluid therapy to optimize cardiac function, sus-
mechanism of action, and toxicities tain tissue perfusion, mitigate organ dysfunction, and achieve
• Understanding of sodium and water physiology physiologic end points. Methods of fluid administration in this
and homeostasis phase vary between bolus, fluid challenge, and continuous
• General knowledge of pathophysiology and infusion. During the stabilization phase, the goals of fluid ther-
differentiation of sodium disorders
apy are to provide ongoing organ support, prevent worsening
• Knowledge of modes of renal replacement therapy organ dysfunction, and avoid further complications. The need
used in critically ill patients
for fluid during this phase is directed toward maintaining intra-
• Knowledge of standard treatment approaches for vascular volume homeostasis and replacing ongoing losses.
sodium abnormalities
Fluid challenge and continuous infusion are the methods of IVF
Table of common laboratory reference values. administration during stabilization. The de-escalation phase
occurs with mobilization and removal of accumulated fluid.
ADDITIONAL READINGS With recovery and decreasing illness severity, a negative fluid
balance is typically targeted and may be achieved by sponta-
The following free resources have additional back- neous or active diuresis of the patient (Rewa 2015).
ground information on this topic:
• Surviving Sepsis Campaign: international guide-
lines for management of sepsis and septic shock: CONTROVERSIES IN FLUID
2016. Crit Care Med 2017;45:486-552. RESUSCITATION
• Verbalis JG, Goldsmith SR, Greenberg A, et al. Crystalloids vs. Colloids
Diagnosis, evaluation, and treatment of
hyponatremia: expert panel recommendations. Theoretical advantages exist for using one type of IVF versus
Am J Med 2013;126:s1-s42. another. Colloids may more effectively maintain intravascu-
lar volume status, leading to potentially less total volume

CCSAP 2018 Book 3 • Fluids and Nutrition/GI and Liver Disorders 8 Fluid and Hyponatremia Management
administered, which may reduce significant edema (Rewa the first ICU week: 2 L (interquartile range [IQR] 1–3.5) com-
2015). However, albumin costs as much as 30 times more pared with 3 L (IQR 0.5–3.2) (p<0.001). Despite equivalent
than normal saline or lactated Ringer solution (Severs 2015). outcomes, this study had many significant limitations. About
Meta-analyses of small studies evaluating the difference 25% of subjects received open-label fluid administration. The
between crystalloid and colloid solutions have had mixed study duration extended over 9 years, which may also reflect
results, leading to questionable effects of colloids on mortal- inconsistencies in practice over that time and changes in
ity (Finfer 2004). guideline recommendations. Finally, the study allowed a vari-
The Saline versus Albumin Fluid Evaluation (SAFE) study ety of fluids to be chosen at the clinician’s discretion, making
was the first large, multicenter, double-blind, controlled trial it extremely difficult to ascertain any difference between the
to evaluate differences in clinical outcomes between crys- specific fluids used. Despite this limitation, the study design
talloid and colloid solutions (Finfer 2004). The SAFE study may make the data somewhat more generalizable because,
randomized 3499 subjects to 4% albumin and 3501 subjects reflecting the reality of clinical practice.
to normal saline. Over the first 2 days, subjects in the albumin Normalizing serum albumin may help maintain oncotic
group received less mean volume than normal saline (1183.9 ± pressure differences that retain volume in the vascula-
973.6 mL on day 1, 602.7 ± 892.7 mL on day 2 vs. 1565.3 ± ture more effectively. This was tested in the Albumin Italian
1536.1 mL on day 1, 954 ± 1484.4 mL on day 2; p<0.001). No Outcome Sepsis (ALBIOS) study, a multicenter, open-label,
significant difference occurred in the primary outcome of randomized controlled trial (Caironi 2014). Patients were ran-
28-day mortality (20.9% albumin vs. 21.1% normal saline, domized to receive crystalloid (n=907) or crystalloid with
p=0.87). Hospital and ICU length of stay (LOS), duration of 20% albumin (n=903) to maintain serum albumin concentra-
mechanical ventilation, need for dialysis, and new organ fail- tions of 3 g/dL. No differences in 28-day mortality occurred
ure were similar. The adjusted OR for death with albumin was between the groups (31.8% albumin vs. 32% crystalloid;
0.71 (95% CI, 0.52–0.97; p=0.03), suggesting its superiority p=0.94), nor were secondary outcomes such as 90-day mor-
to normal saline. However, those who had a traumatic injury tality, ICU/hospital LOS, acute kidney injury (AKI), need for
and received albumin were at higher risk of death than those dialysis, and duration of mechanical ventilation any different.
who received normal saline (13.6% vs. 10%, RR 1.36 [95% CI, However, the albumin group discontinued vasopressors more
0.99–1.86]). Those with traumatic brain injury (TBI) were at quickly than the crystalloid group with a median of 3 days
higher risk of death if they received albumin (24.5% vs. 15.1%, (IQR 1–6) compared with 4 days (IQR 2–7) (p=0.007). Post
p=0.009). A post hoc analysis of the SAFE study, the SAFE- hoc analysis showed a lower 90-day mortality rate in patients
TBI study, corroborated these results, especially in patients with septic shock who received albumin (43.6% vs. 49.9%;
with severe brain injury (Glasgow Coma Scale scores 3–8) p=0.03). Overall, the ALBIOS trial has strengths, including a
(Myburgh 2007). Data from the SAFE study suggests that multicenter design, as well as limitations, including differ-
for overall volume resuscitation in critically ill patients, nor- ences in albumin concentration compared with the SAFE
mal saline and albumin are essentially equivalent in efficacy study and the open-label design. Mortality was less than
and safety. However, albumin solutions should be avoided expected and may have caused the study to be underpow-
in patients with trauma or TBI. In addition, the findings from ered to find differences. However, the data analyses suggest
the subgroup analysis are limited because the study was not that normalization of albumin using exogenous adminis-
designed to evaluate these results. tration does not result in improved mortality but affects
The CRISTAL study was an open-label, multicenter trial that hemodynamic recovery in patients with sepsis. The potential
evaluated outcomes related to the use of crystalloids versus for improved mortality in septic shock requires further study
colloids in patients with hypovolemic shock (Annane 2013). A to confirm these findings.
total of 1414 subjects were in the colloid group, with 1443 in In addition to albumin, hydroxyethyl starch has been com-
the crystalloid group. Subjects in the crystalloid group could pared with crystalloid solutions for resuscitation in critically
receive normal saline or lactated Ringer solution, and those in ill patients. Three large, well-designed studies have shown an
the colloid group could receive albumin (4%, 5%, 20%, or 25%), increased risk of AKI and the need for renal replacement ther-
dextrans, starches, or gelatin solutions. All-cause mortality apy (RRT) using hydroxyethyl starch (Myburgh 2012; Perner
at 28 days was 25.4% in the colloid group compared with 27% 2012; Brunkhorst 2008). In addition, 90-day mortality has
in the crystalloid group (RR 0.96 [95% CI, 0.88–1.04]; p=0.26). been higher in subjects receiving hydroxyethyl starch than in
At 90 days, the crystalloid group had better mortality (number those receiving Ringer acetate (Perner 2012). Hydroxyethyl
needed to treat 29), but the colloid group had more subjects starch has also been associated with an increased risk of
alive not receiving mechanical ventilation or vasopressors. coagulopathy, mainly due to its reduction of factor VIII and
Other secondary outcomes, including dialysis requirement, von Willebrand factor (Severs 2015). Because of an increased
ICU and hospital LOS, and duration of resuscitation therapy, risk of renal injury and mortality, hydroxyethyl starch is not
did not differ between the groups. Subjects in the colloid group recommended for critically ill patients requiring fluid resusci-
received significantly lower median cumulative volumes in tation (Rhodes 2017).

CCSAP 2018 Book 3 • Fluids and Nutrition/GI and Liver Disorders 9 Fluid and Hyponatremia Management
Meta-analyses comparing outcomes between the use of takeaway from the data analyses is that hydroxyethyl starch
crystalloids and colloids have shown mixed results. A 2011 should be avoided in the critically ill population because of
Cochrane review (Roberts 2011) showed that the RR of death the increased risk of renal dysfunction, coagulopathy, and
with colloids was 1.02 (95% CI, 0.92–1.13). The SAFE study mortality.
clearly influenced the results of this meta-analysis because
it composed 75% of the data. Another 2011 meta-analysis Balanced Crystalloid Solutions vs. Normal
showed that using albumin in patients with sepsis reduced Saline Solutions
mortality (OR 0.82 [95% CI, 0.67–1], p=0.047) (Delaney 2011). The most commonly used crystalloids include normal saline
A 2013 Cochrane review that included the use of semisyn- and several balanced crystalloid solutions, including lactated
thetic colloids had a 0.99 RR of death (95% CI, 0.71–1.06) Ringer solution, Plasma-Lyte 148, and Normosol. Table 1
(Perel 2013). details the different composition of some of these IVFs com-
Variability in outcomes, patient populations, and prod- pared with human plasma. Relative to human plasma, normal
ucts used in these clinical trials limits the generalizability saline has a significantly higher sodium and chloride content
of the available data. However, colloid solutions have no and contains no additional electrolytes or buffers.
clear advantage over crystalloids for standard fluid resus- Balanced crystalloid solutions include bicarbonate, lac-
citation in critically ill patients, and fluid choice should be tate, acetate, and a variety of electrolytes. Many of these
patient-specific. These data do not apply to trauma and solutions were developed to be more physiologically compat-
burn patients, in whom specific approaches to fluid man- ible with human plasma. Despite having compositions more
agement have been developed and will not be covered in this similar to plasma, the balanced crystalloid solutions have
chapter. Several guidelines, including those of the Surviving their own set of potential disadvantages. Lactated Ringer
Sepsis Campaign and the National Institute for Health Care solution is hypotonic compared with plasma and may lead to
Excellence, recommend crystalloid solutions over colloids for complications related to cerebral edema in patients with neu-
resuscitating critically ill patients (NICE 2017; Rhodes 2017). rologic emergencies (Ince 2014). Incompatibilities of calcium
This is based on the limited quality of data analyses that sug- exist with other intravenous medications and blood prod-
gest positive outcomes associated with colloid solutions, a ucts, making coadministration with lactated Ringer solution
lack of clear benefit with use of albumin, and the significant difficult. Newer balanced crystalloids like Plasma-Lyte 148
difference in cost between the two fluid types. A combination were developed to overcome some of these negative effects.
of crystalloids and colloids may be used for certain patients Plasma-Lyte 148 and Normosol are both isotonic and may
when large-volume resuscitation is required or to expedite not have the same deleterious effects in patients at risk for
weaning of vasopressor drugs in the setting of shock. Another development or worsening of cerebral edema. The absence

Table 1. Components of Human Plasma and Common Crystalloid Solutions

Lactated
Human Plasma Normal Saline Ringer Solution Normosol-R pH 7.4 Plasma-Lyte 148

Osmolarity (mOsm/L) 275–295 308 273 294 294

pH 7.35–7.45 4.5–7 6–7.5 4–8 7.4

Na (mEq/L) 135–145 154 130 140 140

Cl (mEq/L) 94–111 154 109 98 98

K (mEq/L) 3.5–5.3 0 4 5 5

Calcium (mEq/L) 2.2–2.6 0 2.7 0 0

Magnesium (mEq/L) 0.8–1 0 0 3 1.5

HCO3 (mEq/L) 24–32 0 0 0 0

Acetate (mEq/L) 1 0 0 27 27

Gluconate (mEq/L) 0 0 0 23 23

Lactate (mEq/L) 1–2 0 28 0 0

Information from: Severs D, Hoorn E, Rookmaaker M. A critical appraisal of intravenous fluids: from the physiological basis to clinical
evidence. Nephrol Dial Transplant 2015;30:178-87.

CCSAP 2018 Book 3 • Fluids and Nutrition/GI and Liver Disorders 10 Fluid and Hyponatremia Management
of calcium and lactate may alleviate some of the concerns of population included 760 subjects in the control period and
compatibility and metabolic derangements associated with 773 in the intervention period. Significantly fewer subjects
lactated Ringer solution. in the intervention period developed the primary outcome
A concern of using normal saline in critically ill patients of AKI, as defined by “injury” and “injury and failure” (6.3%
that has garnered significant attention is its association vs. 3%, p=0.002; 14% vs. 8.4%, p<0.001), and fewer subjects
with hyperchloremic acidosis. Classically, the dilutional aci- required RRT. The two groups did not differ in mortality,
dosis theory and the Stewart hypothesis have been used to LOS, or requirement for long-term dialysis. Study limitations
explain hyperchloremic acidosis as a consequence of saline include the single-center, before-after design, which elimi-
administration. Dilutional acidosis occurs with a reduction in nates any direct comparison between IVFs studied and limits
pH, bicarbonate, and anion gap with an increase in chloride the applicability to the ICU that served as the study site. In
concentrations (Doberer 2009). Hyperchloremia has been addition, this was a bundle-of-care study, making it difficult
associated with the development of AKI from sustained renal to ascertain which individual IVF may have caused a differ-
vasoconstriction through A1-receptor-mediated afferent arte- ence in the incidence of AKI.
riole constriction and increased renal intracapsular pressure, The first well-designed prospective study to compare
resulting in decreased tissue perfusion, reduced micro- the effects of normal saline with Plasma-Lyte 148 was
vascular flow, and impaired renal function (Ren 2004). In a the 0.9% saline versus Plasma-Lyte 148 for ICU fluid ther-
retrospective study of 1940 patients with sepsis and septic apy (SPLIT) trial, a prospective, multicenter, randomized,
shock, increases in serum chloride from baseline were asso- blinded, double-crossover controlled trial (Young 2015). A
ciated with significantly increased mortality (adjusted OR total of 1152 subjects received Plasma-Lyte 148, and 1110
1.27; 95% CI, 1.02–1.59; p=0.03) (Neyra 2015). Use of balanced subjects received normal saline. Acute kidney injury did not
crystalloid solutions has reduced hyperchloremic acido- differ in subjects receiving Plasma-Lyte 148 (RR 1.04 [95%
sis in patients undergoing renal transplantation, abdominal CI, 0.80–1.36; p=0.77]). Secondary outcomes did not differ
aortic aneurysm repair, and trauma resuscitation (Potura significantly between the two groups, including RRT, ICU/
2015; Shaw 2012; O’Malley 2005; Waters 2001). A propensi- hospital LOS, duration of mechanical ventilation, and mor-
ty-matched sample of over 30,000 patients requiring open tality. Despite showing no risk of adverse outcomes with
abdominal surgery in a retrospective analysis showed signif- normal saline, this study had several limitations. The mean
icant decreases in postoperative infection, need for dialysis, Acute Physiology and Chronic Health Evaluation II (APACHE
need for blood transfusion, and electrolyte disturbances in II) score was 14.1, and about 70% of subjects were surgical,
those who received Plasma-Lyte 148 compared with normal with 50% having elective procedures. This may account for
saline (Shaw 2012). the population’s relatively low mortality rate (6.6% ICU mor-
In critically ill patients, some retrospective analyses have tality in Plasma-Lyte 148 vs. 7.2% ICU mortality in normal
shown an association between worse outcomes and use of saline and 7.6% hospital mortality in Plasma-Lyte 148 vs.
normal saline. In a sample of 109,836 patients who met sys- 8.6% hospital mortality in normal saline). In addition, AKI
temic inflammatory response syndrome criteria and received was defined as subjects meeting any level of the risk, injury,
crystalloids, mortality was higher (adjusted OR 1.09; 95% CI, failure, loss, and end-stage kidney disease (RIFLE) criteria.
1.06–1.12) with increasing volume-adjusted chloride load Therefore, severity of kidney injury is difficult to elucidate.
(105 mEq/L or more) (Shaw 2014). Another retrospective Chloride concentrations were not collected, so no correla-
study showed that, in a propensity-matched cohort of 6730 tions were possible between hyperchloremia, fluid choice,
patients, use of balanced crystalloids was associated with and AKI. The median total amount of study fluid administered
lower in-hospital mortality than normal saline (RR 0.86; 95% to each group was 2 L, making it difficult to extrapolate the
CI, 0.78–0.94) (Raghunathan 2014). findings to large-volume resuscitation. The study is also lim-
Despite the associations between normal saline and ited because, despite blinding of the investigators to study
worse clinical outcomes, including mortality, these studies fluid, two-thirds of the investigators were able to accurately
had small sample sizes or were retrospective and had popu- identify the Plasma-Lyte 148 administered. That open-label
lations with significant heterogeneity. The first prospective Plasma-Lyte 148 was allowed according to the protocol intro-
study to evaluate the effects of various crystalloid solutions duces potential for bias.
in critically ill patients was conducted in a single center in The Isotonic Solutions and Major Adverse Renal Events Trial
Australia (Yunos 2012). This was an open-label, before-after (SMART) was a single-center, unblinded, cluster-randomized,
study in a 22-bed ICU. The study allowed use of normal multiple-crossover trial that compared normal saline with
saline, 4% succinylated gelatin solution, and 4% albumin in balanced crystalloid solutions (Semler 2018). A total of 7942
the control period (chloride-liberal), followed by a phase-out subjects were randomized to receive balanced crystalloids,
period, and then the intervention period (chloride-restrictive), with 7860 to receive normal saline. One major difference of
in which clinicians could choose between lactated Ringer the SMART trial compared with the SPLIT trial is that in the
solution, Plasma-Lyte 148, and 20% albumin. The study SMART trial, the balanced crystalloid group could receive

CCSAP 2018 Book 3 • Fluids and Nutrition/GI and Liver Disorders 11 Fluid and Hyponatremia Management
either Plasma-Lyte 148 or lactated Ringer solution. A signif- receiving normal saline (Self 2018). The median volume of
icant difference in the primary outcome of major adverse study fluid administered in each group was 1 L (IQR 1–2).
kidney events in 30 days was found between the two groups Although hospital-free days did not differ significantly (up
(14.3% in balanced crystalloid group vs. 15.4% normal saline to 28 days), subjects receiving balanced crystalloids had a
group; adjusted OR 0.9 [95% CI, 0.82–0.99], p=0.04). In addi- lower frequency of an adverse kidney event (4.7% vs. 5.6%,
tion, subjects who received balanced crystalloids had a OR 0.82 [95% CI, 0.7–0.95], p=0.01). Although the SALT-ED
higher mean number of RRT-free days (25 ± 8.6 vs. 24.8 ± 8.9, trial was specifically of non-critically ill patients, the find-
OR 1.11 [95% CI, 1.02–1.2]). Although mortality did not dif- ings are consistent because the risk of AKI appears to be
fer significantly, in-hospital death before 30 days trended in increased using normal saline compared with balanced
favor of the balanced crystalloid group (10.3% vs. 11.1%, OR crystalloids.
0.9 [95% CI, 0.8–1.01], p=0.06). Subgroup analysis revealed According to the available data, normal saline may increase
significant decreases in the composite outcome of death, the risk of hyperchloremic metabolic acidosis and AKI, even
new RRT, or persistent renal dysfunction in subjects in the at the low cumulative volumes administered. The associ-
medical and neurologic ICUs, subjects with sepsis, subjects ation between normal saline use and increased mortality
without TBI, and those who were receiving RRT before study deserves further study because several retrospective analy-
enrollment. Chloride concentrations were collected and ses have indicated such a relationship. Data from the SMART
were similar at baseline (median 103 mEq/L [IQR 100–106]) trial showing a potential mortality benefit of using balanced
in the balanced crystalloid group compared with 103 mEq/L crystalloid solutions are merely speculative because the
(IQR 100–106 in the normal saline group). However, signifi- study was not designed to find a difference in that outcome.
cantly more subjects had a Cl concentration greater than 110 Currently, clinical practice guidelines do not recommend
mEq/L in the normal saline group (24.5% vs. 35.6%, p<0.001) use of balanced crystalloid solutions over normal saline,
and an HCO3 less than 20 mEq/L (35.2% vs. 42.1%, p<0.001). aside from avoidance of hyperchloremia (NICE 2017; Rhodes
Ventilator-free, ICU-free, and vasopressor-free days were not 2017). Recent publication of the SMART and SALT-ED stud-
significantly different between the two groups. The SMART ies may alter these guideline recommendations in the future.
trial has many strengths and limitations. Overall, the trial However, additional data are required in critically ill patients
was well designed and had a large sample size. Chloride at high risk of death and those who receive large volumes of
concentrations were collected and correlated with the fluid IVF, greater than 60 mL/kg in a 24-hour period (Sen 2017).
type used. The study population differed from that in the In addition, it is unclear whether Plasma-Lyte 148 is superior
SPLIT trial, with 21.4% being surgical admissions and 14.8% to lactated Ringer solution, which is more readily available
having a diagnosis of sepsis or septic shock (3.4% of sub- and less costly. The average wholesale cost of Plasma-
jects had sepsis in the SPLIT trial). The mean cumulative Lyte 148 is about 5 times that of normal saline (Smith 2014).
volume of study fluid administered from ICU admission to Currently, it seems prudent to carefully select a crystalloid
hospital discharge or 30 days was 2.3 plus or minus 3.6 L fluid on the basis of patient characteristics and institutional
in the balanced crystalloid group and 2.6 plus or minus 4.5 availability. Although data analyses suggest increased risk
L in the normal saline group. Similar to the SPLIT trial, this in harm with normal saline, further study is required to con-
makes it difficult to extrapolate the SMART study’s findings firm these findings. In the meantime, lactated Ringer solution
to large-volume resuscitation. The estimated in-hospital may be considered an optimal fluid choice in most critically ill
mortality was 9.4% (95% CI, 9–9.9) in the balanced crystal- patients requiring fluid administration, with Plasma-Lyte 148
loid group compared with 9.6% (95% CI, 9.2–10) in the normal and Normosol as alternative options because of their high
saline group. This is difficult to compare with the SPLIT trial cost. According to the available data, balanced crystalloid
because the predicted mortality was estimated using the solutions are preferred in patients who receive large-volume
Vizient database and cannot reliably be related to use of resuscitation, have or develop hyperchloremia, or are receiv-
the APACHE II score. Other limitations include the study’s ing RRT.
single-center and unblinded design. Finally, the risk of AKI
between use of lactated Ringer solution and Plasma-Lyte
UPDATES IN MANAGING
148 cannot be distinguished in this study because the bal-
HYPONATREMIA IN THE ICU
anced crystalloid group could receive either fluid.
Recent data analyses have had similar findings in the Introduction
non-critically ill population when comparing normal saline Hyponatremia may cause complications in critically ill patients,
with balanced crystalloid solutions. The Saline against including neurologic deficits, muscle weakness and cramps,
Lactated Ringer’s or Plasma-Lyte 148 in the Emergency hyperventilation, impairment in gluconeogenesis, and decreased
Department (SALT-ED) trial was a single-center, unblinded, left ventricular function (Verbalis 2013). The presence of hypo-
multiple-crossover trial that included 6708 subjects receiv- natremia in the ICU is associated with mortality rates as high as
ing Plasma-Lyte 148 or lactated Ringer solution and 6639 37.7%. Management of hyponatremia depends on its underlying

CCSAP 2018 Book 3 • Fluids and Nutrition/GI and Liver Disorders 12 Fluid and Hyponatremia Management
Patient Care Scenario
A 63-year-old woman (weight 63 kg) with a history of mechanical ventilation and sedated. Laboratory values
chronic obstructive pulmonary disease and heart failure are within normal limits, except for Cl 108 mEq/L. Given
(ejection fraction of around 35%) was admitted for sep- the patient’s presentation, develop a plan to assess for
tic shock secondary to a UTI. She has received 2 L of fluid responsiveness, what type of fluid to administer, and
normal saline (bolus) and is receiving a norepinephrine how to monitor therapy.
infusion for shock at 30 mcg/minute. She is intubated on

ANSWER
Fluid responsiveness should be reassessed often because because she is at higher risk of developing pulmonary
the patient is now receiving vasopressor therapy. edema, given her underlying heart failure. Additional crys-
Evaluation of the patient’s intravascular volume should talloid resuscitation may exacerbate this negative effect.
begin with a physical examination, including capillary Therefore, it may be reasonable to administer 5% albumin
refill, skin turgor, extremity temperature, mucus mem- 250–500 mL instead of saline, followed by reassessment
branes, urine color, and urinary output. Laboratory tests of intravascular volume status. Use of albumin has been
such as SCr, BUN, and lactate may also indicate fluid sta- associated with faster resolution of hemodynamic instabil-
tus. Dynamic assessments should be used over static ity when combined with crystalloids. Because this patient
measurements because dynamic assessments tend to is taking relatively high norepinephrine doses, albumin in
be more reliable measurements of intravascular volume addition to crystalloid resuscitation may be of benefit.
responsiveness. Passive leg raises and pulse pressure Hyperchloremic metabolic acidosis has been asso-
variation are noninvasive and easily done to assess fluid ciated with 0.9% sodium chloride, and data analyses
responsiveness. If the equipment is available, the clini- suggest increased risk of renal dysfunction using this
cian may do esophageal Doppler monitoring or ultrasound fluid. Because the patient has already developed hyper-
assessment of the IVC. chloremia, lactated Ringer solution would be optimal if
Regarding selecting the best fluid for this patient, no crystalloid resuscitation were desired. In addition, no data
clear evidence shows that colloid solutions are superior suggest a benefit of Plasma-Lyte or Normosol over lac-
to crystalloids across the board. Albumin may result in tated Ringer solution; thus, the cost of these agents would
less total volume administered and has clinical outcomes not make them first line. Additional monitoring of chloride
similar to saline. This may be consequential in the patient and renal function is recommended.

1. A nnane D, Siami S, Jaber S, et al. Effects of fluid resuscitation with colloids vs crystalloids on mortality in critically ill patients present-
ing with hypovolemic shock: the CRISTAL randomized trial. JAMA 2013;310:1809-17.
2. Semler MW, Self WH, Wanderer JP, et al. Balanced crystalloids versus saline in critically ill adults. N Engl J Med 2018;378:829-39.

cause and may include infusion of normal saline, hypertonic to place a central venous catheter. This may be unaccept-
saline, and water restriction (for syndrome of inappropriate able in patients with severe TBI or hyponatremia who require
antidiuretic hormone [SIADH]). Several limitations to widely urgent administration of hypertonic saline. At least three ret-
accepted treatment modalities exist, including unpredictable rospective studies have evaluated the safety of peripheral
and rapid changes in sodium and administration-related compli- administration of 3% sodium chloride (Table 2).
cations of hypertonic saline. All three studies suggest that peripheral administration of
3% saline is a safe alternative in managing hyponatremia, espe-
Peripheral Administration of Hypertonic Saline cially in patients for whom timely placement of central venous
Solutions for Hyponatremia access is not achievable. The frequency of infusion-related
Hypertonic saline is typically reserved for patients with reactions such as phlebitis and extravasation was low across
severe hyponatremia (Na less than 120 mEq/L or with asso- all the studies. One subject developed brachial vein throm-
ciated symptoms). Hypertonic saline is also commonly used bosis without further complications. In the largest study,
in patients with intracranial hypertension secondary to TBI phlebitis or extravasation occurred in about 7% of the popula-
or other causes of cerebral edema. One of the main concerns tion. Of note, this study included two separate centers, where
regarding use of hypertonic saline is the need for appropriate one institution had a maximum rate of 3% sodium chloride of
administration to minimize complications related to infusion 30 mL/hour and the other, 75 mL/hour. On further analysis, 8
of the fluid. Historically, 3% sodium chloride solutions are of the 15 subjects who developed an infusion-related adverse
administered through a central venous catheter because of event had their infusion changed to an alternative peripheral
the risk of phlebitis and extravasation-related injuries from site, 5 had their infusion changed to central administration,
its hyperosmolarity (Dillon 2018). Central venous catheters and 2 had their infusion completely stopped. No additional
are associated with an increased risk of infection, pneumo- complications were noted in the patients who had peripheral
thorax, and thrombosis. In addition, significant delays in administration continued through an alternative site. Of note,
effective therapy may occur because of the time required most subjects had at least a 20-gauge catheter in place, and

CCSAP 2018 Book 3 • Fluids and Nutrition/GI and Liver Disorders 13 Fluid and Hyponatremia Management
Table 2. Studies Evaluating Peripheral Administration of 3% NaCl

Dillon RC 2018 Perez CA 2017 Jones GM 2016

Sample size 66 28 213

Indication
Hyponatremia 42 (64%) 0 (0%) 0 (0%)
Neurologic injury 19 (29%) 28 (100%) 213 (100%)

Catheter gauge
≤ 20 101 (60%) 28 (100%) 208 (97.7%)
≥ 22 67 (40%) 0 (0%) 5 (2.3%)

Infusion rate (mL/hr)

Initial 30 (IQR 25–43) Not described 30 (IQR 20–30)
Median/mean 34 (IQR 30–42) 39 ± 10 30 (IQR 24.4–34.7)
Maximum 50 (IQR 8–75) 50 30 (IQR 25–40)
Range 30–50

Median/mean infusion duration (hr) 14 (IQR 4–30) 36 (range 1–124) 0.85 (IQR 0.44–1.36)

Infusion-related adverse events

• Phlebitis 2 1 9

• Extravasation 0 0 6

• Venous thrombosis 0 1 0

NaCl = sodium chloride.

Information from: Dillon RC, Merchan C, Altshuler D, et al. Incidence of adverse events during peripheral administration of sodium
chloride 3%. J Intensive Care Med 2018;33:48-53; Perez CA, Figueroa SA. Complication rates of 3% hypertonic saline infusion
through peripheral intravenous access. J Neurosci Nurs 2017;49:191-5; and Jones GM, Bode L, Riha H, et al. Safety of continuous
peripheral infusion of 3% sodium chloride solution in neurocritical care patients. Am J Crit Care 2016;26:37-42.

the infusion rates were mainly 30–40 mL/hour. Therefore, it have delays in establishing central venous access. However,
seems prudent to ensure that larger catheters are used and to additional data from larger and/or well-designed controlled
maintain low infusion rates if the peripheral route is chosen. trials are required to ensure the safety of this treatment
Data from the largest study are limited because of the short approach.
infusion duration. However, the other two studies suggest
that infusions beyond 1 hour (up to 24–36 hours) can safely Adjunctive Desmopressin to Prevent Sodium
be administered through a peripheral line. Despite promis- Overcorrection in Hyponatremia
ing results from these three studies, small sample size and Sodium overcorrection may lead to devastating neurologic
retrospective design significantly limit their impact on prac- consequences. Current expert panel recommendations
tice, and additional larger, prospective studies are required include limiting the increase in sodium to 8 mEq/L in 24 hours
to confirm these findings before peripheral administration of for patients at high risk of osmotic demyelination and to
hypertonic saline can routinely be recommended. 10–12 mEq/L in 24 hours and 18 mEq/L in 48 hours in those
Another study compared the efficacy and safety of 5% and at normal risk (Verbalis 2013). Those at high risk of osmotic
23.4% sodium chloride in subjects with intracranial hyperten- demyelination include patients with Na less than 105 mEq/L,
sion (Carter 2017). The study was conducted retrospectively hypokalemia, alcoholism or advanced liver disease, and mal-
at a single institution and included data from 11 patients nutrition. Hypertonic saline solutions may increase the risk
who received 5% sodium chloride administered peripherally. of osmotic demyelination because they can replete sodium
The infusion according to institutional protocol was admin- more rapidly. One strategy to limit sodium overcorrection is
istered over 15 minutes. No infusion-related adverse events using desmopressin in patients at high risk. Desmopressin
were documented. Data analyses from this study suggest is a synthetic analog of antidiuretic hormone that binds to
that in patients who require rapid lowering of intracranial vasopressin 2 (V2) receptors in the kidney, increasing free
pressure, peripheral administration of 5% sodium chloride is water reabsorption through aquaporin channel translocation
an alternative to 23.4% solutions, especially in patients who (Rafat 2014). Theoretically, desmopressin can decrease the

CCSAP 2018 Book 3 • Fluids and Nutrition/GI and Liver Disorders 14 Fluid and Hyponatremia Management
rapid rise in serum sodium as patients are actively treated unclear. This strategy should be combined with replacing free
for hyponatremia. Three retrospective studies have evalu- water losses with 5% dextrose or oral water. Once the sodium
ated desmopressin to prevent overcorrection in patients with reaches 128 mEq/L, desmopressin should be discontinued.
severe hyponatremia (Na less than 120 mEq/L). One study In patients requiring re-lowering of sodium in the setting of
of 25 subjects evaluated the effects of desmopressin 1–2 overcorrection, desmopressin can be administered with free
mcg intravenously or subcutaneously every 6–8 hours (Sood water replacement at 3 mL/kg/hour (Verbalis 2013).
2013). No subjects had an increase greater than 12 mEq/L in
24 hours or greater than 18 mEq/L in 48 hours. No subject Vasopressin Antagonists for Hyponatremia
was identified as having neurologic complications or a com- in Critically Ill Patients
plication related to desmopressin administration. Another drug class that has been used for hyponatremia in
Another study reported the effects of desmopressin in critically ill patients is vasopressin antagonists (“vaptans”).
patients who had a sodium correction greater than 12 mEq/L These agents are high-affinity nonpeptide antagonists of argi-
in 24 hours (n=6) or those at risk (n=14) (Perianayagam 2008). nine vasopressin V2 and V1A receptors. Vaptans produce renal
Five subjects received hypertonic saline, three of whom excretion of solute-free water, sparing sodium (Lehrich 2012).
received 1- to 2-mcg doses of desmopressin in the study. In The available agents are conivaptan and tolvaptan. According
subjects who had overcorrection in the first 48 hours, only to the labeling, conivaptan is administered intravenously with
one had a sodium increase greater than 18 mEq/L in 48 hours. a loading dose of 20 mg over 30 minutes, followed by 20 mg as
None of the subjects at risk had sodium increases beyond a continuous infusion over 24 hours, which can be continued
these limitations at 24 and 48 hours. Similar to the previ- up to 4 days. Tolvaptan is an oral agent with a starting dose
ous study, none of the subjects had significant neurologic of 15 mg/day that can be titrated to 60 mg daily, depending
complications. on serum sodium response (Lehrich 2012). Both conivaptan
The third study was another retrospective analysis in two and tolvaptan are substrates of CYP3A4 and are contrain-
ICUs including 20 subjects (Rafat 2014). Subjects in this study dicated in the setting of using strong CYP3A4 inhibitors.
received doses of 2–4 mcg intravenously of desmopressin. Tolvaptan is preferred in patients with cirrhosis because it is
The sodium correction rate was significantly decreased after more selective to V2 receptors. Conivaptan is a nonselective
receiving desmopressin (-0.02 mEq/L/hour vs. 0.81 mEq/L/ vasopressin antagonist that may lead to decreased renal per-
hour; p=0.001). Eleven subjects required a re-lowering of fusion and hepatorenal syndrome. The vaptans are relatively
sodium. So many cases requiring re-lowering indicates that well tolerated, with infusion-site reactions being the most
the general approach to management was flawed in the study common adverse effect with conivaptan, together with hypo-
sites. Of note, only three subjects were treated using hyper- kalemia, orthostatic hypotension, and fever (Lehrich 2012).
tonic saline, and several patients received hypotonic fluids. These agents are labeled to treat euvolemic or hypervolemic
Because of its retrospective nature and incomplete reporting hyponatremia and can correct sodium in patients with heart
of how the study sites managed hyponatremia, these find- failure. However, vaptans should not be used to treat hypo-
ings are difficult to interpret. None of the patients receiving volemic hyponatremia. In addition, free water restrictions
desmopressin had seizures; however, one patient developed should be discontinued when vaptans are used to prevent sig-
central pontine myelinosis. This patient’s sodium increased nificant volume losses (Lehrich 2012). Compared with other
by 10 and 16 mEq/L at 24 and 48 hours, respectively. Of methods of treating hyponatremia such as hypertonic saline
interest, this patient did not receive hypertonic saline. He pre- and desmopressin, vaptans are significantly more costly.
sented with hypokalemia and had a history of alcoholism, Nevertheless, vaptans may be an attractive alternative in
placing him at high risk of osmotic demyelination. treating patients with hyponatremia for a variety of reasons.
The main limitations to these data include retrospective For example, conventional therapies like hypertonic saline do
designs, small sample sizes, and lack of ability to effec- not always produce a predictable change in sodium and may
tively measure effects on neurologic outcomes. Because of increase the risk of osmotic demyelination (Murphy 2009). As
these limitations, it is unclear whether desmopressin alone discussed earlier, evidence showing that desmopressin may
was responsible for the changes in sodium versus careful prevent sodium overcorrection is derived entirely from ret-
initiation and close monitoring of hypertonic saline admin- rospective analyses. In addition, the debate of using central
istration. However, this strategy appears safe and effective versus peripheral access with hypertonic saline is ongoing.
in the absence of well-designed studies. Expert panel rec- Finally, using volume restriction in patients with SIADH may
ommendations include the adjunctive use of 2- to 4-mcg result in intravascular depletion and decreased cerebral per-
doses of desmopressin to prevent sodium overcorrection in fusion (Murphy 2009).
patients with severe hyponatremia (Verbalis 2013). Despite Two prospective studies have compared conivaptan and
the recommended doses by the expert panel, two of the stud- hypertonic saline for treating SIADH in critically ill patients.
ies previously described used 1- to 2-mcg doses. Therefore, The first was a prospective, single-center study that included
the exact dose used to prevent overcorrection remains patients with SIADH without head injury or neurosurgery

CCSAP 2018 Book 3 • Fluids and Nutrition/GI and Liver Disorders 15 Fluid and Hyponatremia Management
Table 3. Change in Sodium and Daily Fluid Balance Comparison

Time Conivaptan 3% NaCl p Value

Sodium (mEq/L) 12 hr 129.4 ± 2.8 127.7 ± 1.1 0.006

24 hr 131.5 ± 2.1 127.3 ± 2.8 < 0.001

48 hr 130.5 ± 2.6 126.5 ± 1.9 < 0.001

72 hr 129.4 ± 2.5 127.5 ± 3.8 0.159

Daily fluid balance (mL) Day 1 −1094.3 ± 286.7 −425 ± 165 < 0.001

Day 2 −579 ± 375.8 −260.5 ± 282.5 0.005

Day 3 −370 ± 219.9 +263.5 ± 546 < 0.001

Information from: Reddy SN, Rangappa P, Jacob I, et al. Efficacy of conivaptan and hypertonic (3%) saline in treating hyponatremia due
to syndrome of inappropriate antidiuretic hormone in a tertiary intensive care unit. Indian J Crit Care Med 2016;20:714-8.

(Reddy 2016). Eighty subjects were included, with 40 receiv- of conivaptan differs from how it is typically administered.
ing conivaptan and 40 receiving 3% sodium chloride. Each Baseline characteristics were similar between the groups,
treatment was administered until serum sodium reached 130 including average (SD) sodium of 125.5 plus or minus 2.4
mEq/L or up to 72 hours, whichever came first. Most subjects mEq/L in the conivaptan group and 123.9 plus or minus 2.9
were older than 75 and had idiopathic etiology of SIADH. mEq/L in the hypertonic saline group. Table 3 contains the
The average (SD) baseline sodium was 112.8 plus or minus results comparing the two therapies. Conivaptan resulted in
4 mEq/L in the 3% sodium chloride group compared with 114 statistically significantly higher sodium concentrations for
plus or minus 6.4 mEq/L in the conivaptan group. Baseline the first 48 hours and a more negative fluid balance during
characteristics did not differ significantly, and the predicted the 3 days postdose than did 3% sodium chloride. Study lim-
mortality in each group was 26%. No significant differences itations include small sample size and restricted patient
between groups occurred in sodium at 6, 12, and 24 hours. population. Although the study had statistical differences, it
However, at 48 and 72 hours, the conivaptan group had sig- is unclear whether the difference in sodium is clinically sig-
nificantly higher average (SD) serum sodium (128.9 ± 2.6 nificant between use of conivaptan and hypertonic saline.
mEq/L vs. 133 ± 3.8 mEq/L, p<0.001, and 133.7 ± 1.2 mEq/L vs. However, the difference in fluid balance appears to be clini-
135.9 ± 1.4 mEq/L, p<0.001, respectively). The average time cally significant and may affect clinical outcomes. This study
to achieve an Na concentration greater than 130 mEq/L was shows that conivaptan as a single bolus may be at least as
significantly shorter in subjects receiving conivaptan than in effective in increasing sodium at a reasonable rate as hyper-
those receiving 3% sodium chloride (54.60 ± 13.30 hours vs. tonic saline while maintaining a greater negative fluid balance.
66.15 ± 13.29 hours, p<0.001). The LOS values in the ICU and Additional findings from the neurocritical care literature
hospital were also longer in patients who received 3% sodium may support the use of bolus doses of conivaptan. Euvolemic
chloride (ICU 4.61 ± 0.91 days vs. 3.35 ± 0.89 days; p<0.001; hyponatremia is common in patients with neurologic emer-
hospital 6.41 ± 1.41 days vs. 5.65 ± 1.45 days; p<0.001). The gencies and usually develops from SIADH (Marik 2013).
finding of shorter stays in the ICU and hospital is intriguing Development of hyponatremia in these patients may result
because this may offset the higher cost of using conivap- in cerebral edema, herniation, and mortality. Historically,
tan to treat SIADH. The study is limited because of its small hypertonic saline has been used to treat hyponatremia in this
sample size and unblinded design. Therefore, these findings patient population. However, several retrospective studies
require further study before any definitive conclusions can have shown that a single 20-mg intravenous dose of conivap-
be made regarding conivaptan use over traditional methods tan raises serum sodium with minimal adverse effects (Marik
such as hypertonic saline solutions. 2013; Human 2012; Murphy 2009). Some of these studies
The second study was a prospective, randomized trial of also had patients who received several conivaptan doses at
patients who developed hyponatremia postoperatively after doses of 10 or 40 mg. Therefore, it is difficult to make a defin-
head and neck surgery (Rajan 2015). The two treatment arms itive recommendation of whether single or multiple doses of
consisted of 20 subjects who received a single 20-mg intra- conivaptan are optimal for these patients. However, a single
venous bolus of conivaptan and 20 subjects who received bolus dose may decrease the cost of using conivaptan and
a 3% sodium chloride infusion at 20–30 mL/hour. The inter- could limit the chance for volume overload. In addition, a con-
vention is notable because the use of a single-bolus dose tinuous infusion of conivaptan requires a dedicated catheter,

CCSAP 2018 Book 3 • Fluids and Nutrition/GI and Liver Disorders 16 Fluid and Hyponatremia Management
which may affect administration of other drugs because of
Practice Points
compatibility. Of importance, these suggestions have not
been validated and serve only as potential outcomes for Despite several well-designed clinical trials evaluating
IVF therapy in critically ill patients, many questions
future research.
remain. In addition, various treatment approaches have
Despite the promising results from several studies, current been described for ICU patients with electrolyte disorders:
recommendations do not support use of vaptans in treating • Crystalloid solutions are preferred for most patients who
severe hyponatremia (Na less than 120 mEq/L). Hypertonic require fluid resuscitation because of their equal efficacy
saline solutions are still recommended as first line for this and lower cost compared with colloid solutions. Colloid
indication (Verbalis 2013). However, vaptans can be con- solutions may be considered in patients who may be
limited in total volume administration or to enhance the
sidered as an alternative to hypertonic saline, especially in
efficacy of crystalloid resuscitation.
patients with hyponatremia/SIADH in the setting of acute
• Hydroxyethyl starch should not be used for fluid resuscita-
neurologic conditions such as intracerebral or subarach- tion in critically ill patients.
noid hemorrhage, or in those who require limited volume • Balanced crystalloids such as lactated Ringer solution,
administration. Plasma-Lyte 148, and Normosol may be considered as
alternatives to normal saline to prevent hyperchloremic
metabolic acidosis. Normal saline may increase the risk of
CONCLUSION
AKI in patients requiring fluid therapy.
Despite widespread IVF use, the true efficacy and safety of • Despite historical use of administering 3% sodium chloride
IVFs in the critically ill population remains unknown. Newer through central venous access, peripheral administration
data suggest that for most patients, crystalloid solutions may be relatively safe when given for short periods. This
method of administration may be considered if timely
such as normal saline or lactated Ringer solution are pre-
administration of hypertonic saline is required in patients
ferred. However, colloid solutions, specifically albumin, play with severe hyponatremia.
a role in large-volume resuscitation and in restoring hemody- • Desmopressin administered with hypertonic saline may
namic stability in patients with septic shock. Additional data limit the rate of sodium increase in patients with severe
analyses have shown an increased risk of hyperchloremic hyponatremia. This may provide a useful tool to prevent
metabolic acidosis and renal dysfunction associated with devastating adverse outcomes of sodium overcorrection.

the use of the prototypical crystalloid solution, normal saline.

Use of balanced crystalloid solutions (e.g., lactated Ringer
solution, Plasma-Lyte 148, or Normosol) may reduce the risk REFERENCES
of these adverse outcomes. Currently, it is unclear whether Annane D, Siami S, Jaber S, et al. Effects of fluid resuscita-
major clinical outcomes such as mortality are affected by the tion with colloids vs crystalloids on mortality in critically
use of balanced crystalloids. ill patients presenting with hypovolemic shock: the
Electrolyte disturbances in the critically ill population are CRISTAL randomized trial. JAMA 2013;310:1809-17.
common and require astute recognition, prompt treatment, Brunkhorst FM, Engel C, Bloos F, et al. Intensive insulin ther-
and effective prevention. Traditional methods of managing apy and pentastarch resuscitation in severe sepsis. N Engl
hyponatremia include the requirement of establishing central J Med 2008;358:125-39.
venous access to administer hypertonic solutions. Recent
Caironi P, Tognoni G, Masson S, et al. Albumin replacement
data analyses suggest that for short durations, peripheral
in patients with severe sepsis or septic shock. N Engl J
administration is a reasonable alternative and is helpful for Med 2014;370:1412-21.
patients in whom establishing invasive vascular access is
Carter C, Human T. Efficacy, safety, and timing of 5% sodium
difficult. Treatment of hyponatremia is also associated with
chloride compared with 23.4% sodium chloride for
significant CNS abnormalities in the setting of overcorrection.
osmotic therapy. Ann Pharmacother 2017;51:625-9.
Desmopressin may be a useful adjunctive therapy to prevent
the development of sodium overcorrection in those at risk. Delaney AP, Dan A, McCaffrey J, et al. The role of albumin as
In addition, vasopressin antagonists have been more specif- a resuscitation fluid for patients with sepsis: a systematic
review and meta-analysis. Crit Care Med 2011;39:386-91.
ically studied in the critically ill population with SIADH and
may be reasonable alternatives for patients with this electro- Dillon RC, Merchan C, Altshuler D, et al. Incidence of adverse
lyte disorder. Use of bolus conivaptan is intriguing because events during peripheral administration of sodium
it may prevent the need for central line placement for use of chloride 3%. J Intensive Care Med 2018;33:48-53.
hypertonic saline, can limit the volume of drug/fluid infused, Doberer D, Funk GC, Kirchner K, et al. A critique of Stewart’s
and may mitigate the cost of therapy using a single dose. approach: the chemical mechanism of dilutional acidosis.
Appropriate selection of fluids or drugs and monitoring for Intensive Care Med 2009;35:2173-80.
safety and efficacy are important ways for the critical care Finfer S, Bellomo R, Boyce N, et al. A comparison of albumin
pharmacist to contribute to the care of patients requiring IVF and saline for fluid resuscitation in the intensive care unit.
or management of electrolyte disorders. N Engl J Med 2004;350:2247-56.

CCSAP 2018 Book 3 • Fluids and Nutrition/GI and Liver Disorders 17 Fluid and Hyponatremia Management
Hoste EA, Maitland K, Brudney CS, et al. Four phases of intra- end-stage renal disease undergoing cadaveric renal trans-
venous fluid therapy: a conceptual model. Br J Anaesth plantation: a prospective randomized controlled trial.
2014;113:740-7. Anesth Analg 2015;120:123-9.

Human T, Onuoha A, Diringer M, et al. Response to a bolus of Rafat C, Schortgen F, Gaudry S. Use of desmopressin acetate
conivaptan in patients with acute hyponatremia after brain in severe hyponatremia in the intensive care unit. Clin J
injury. J Crit Care 2012;27:745.e1-5. Am Soc Nephrol 2014;9:229-37.

Ince C, Groeneveld AB. The case for 0.9% NaCl: is the unde- Raghunathan K, Shaw A, Nathanson B, et al. Association
fendable, defensible? Kidney Int 2014;86:1087-95. between the choice of IV crystalloid and in-hospital
mortality among critically ill adults with sepsis. Crit Care
Kalantari K, Chang JN, Ronco C, et al. Assessment of intra- Med 2014;42:1585-91.
vascular volume status and volume responsiveness in
critically ill patients. Kidney Int 2013;83:1017-28. Rajan S, Srikumar S, Paul J, et al. Effectiveness of single
dose conivaptan for correction of hyponatraemia in
Lehrich RW, Greenberg A. Hyponatremia and the use of post-operative patients following major head and neck
vasopressin receptor antagonists in critically ill patients. surgeries. Indian J Anaesth 2015;59:416-20.
J Intensive Care Med 2012;27:207-18.
Reddy SN, Rangappa P, Jacob I, et al. Efficacy of conivap-
Malbrain M, Marik PE, Witters I, et al. Fluid overload, tan and hypertonic (3%) saline in treating hyponatremia
de-resuscitation, and outcomes in critically ill or injured due to syndrome of inappropriate antidiuretic hormone
patients: a systematic review with suggestions for clinical in a tertiary intensive care unit. Indian J Crit Care Med
practice. Anaesthesiol Intensive Ther 2014;46:361-80. 2016;20:714-8.
Marik PE, Rivera R. Therapeutic effect of conivaptan Ren Y, Garvin JL, Liu R, et al. Role of macula densa adenos-
bolus dosing in hyponatremic neurosurgical patients. ine triphosphate (ATP) in tubuloglomerular feedback.
Pharmacotherapy 2013;33:51-5. Kidney Int 2004;66:1479-85.
Murphy T, Dhar R, Diringer M. Conivaptan bolus dosing for Rewa O, Bagshaw SM. Principles of fluid management.
the correction of hyponatremia in the neurointensive care Crit Care Clin 2015;31:785-801.
unit. Neurocrit Care 2009;11:14-9.
Rhodes A, Evans LE, Alhazzani W, et al. Surviving Sepsis
Myburgh J, Cooper DJ, Finfer S, et al. Saline or albumin for Campaign: international guidelines for manage-
fluid resuscitation in patients with traumatic brain injury. ment of sepsis and septic shock: 2016. Crit Care Med
N Engl J Med 2007;357:874-84. 2017;45:486-552.
Myburgh JA, Finfer S, Bellomo R, et al. Hydroxyethyl starch Roberts I, Blackhall K, Alderson P, et al. Human albumin solu-
or saline for fluid resuscitation in intensive care. N Engl J tion for resuscitation and volume expansion in critically ill
Med 2012;367:1901-11. patients. Cochrane Database Syst Rev 2011;11:CD001208.
National Institute for Health and Care Excellence (NICE). Self WH, Semler MW, Wanderer JP, et al. Balanced crystal-
Intravenous Fluid Therapy: Intravenous Fluid Therapy in loids versus saline in non-critically ill adults. N Engl J Med
Adults in Hospital. NICE Clinical Guideline 174. 2017. 2018;378:819-28.
Neyra JA, Canepa-Escaro F, Li X, et al. Association of hyper- Semler MW, Self WH, Wanderer JP, et al. Balanced crystal-
chloremia with hospital mortality in critically ill septic loids versus saline in critically ill adults. N Engl J Med
patients. Crit Care Med 2015;43:1938-44. 2018;378:829-39.
O’Malley CM, Frumento RJ, Hardy MA, et al. A randomized, Sen A, Keener CM, Sileanu FE, et al. Chloride content of flu-
double-blind comparison of lactated Ringer’s solution ids used for large-volume resuscitation is associated with
and 0.9% NaCl during renal transplantation. Anesth Analg reduced survival. Crit Care Med 2017;45:e146-e153.
2005;100:1518-24.
Severs D, Hoorn E, Rookmaaker M. A critical appraisal of
Perel P, Roberts I, Ker K. Colloids versus crystalloids for fluid intravenous fluids: from the physiological basis to clinical
resuscitation in critically ill patients. Cochrane Database evidence. Nephrol Dial Transplant 2015;30:178-87.
Syst Rev 2013;2:CD000567.
Shaw AD, Bagshaw SM, Goldstein SL, et al. Major compli-
Perianayagam A, Sterns RH, Silver SM, et al. DDAVP is effec- cations, mortality, and resource utilization after open
tive in preventing and reversing inadvertent overcorrection abdominal surgery: 0.9% saline compared to Plasma-Lyte.
of hyponatremia. Clin J Am Soc Nephrol 2008;3:331-6. Ann Surg 2012;255:821-9.
Perner A, Haase N, Guttormsen AB, et al. Hydroxyethyl Shaw AD, Raghunathan K, Peyerl FW, et al. Association
starch 130/0.42 versus Ringer’s acetate in severe sepsis. between intravenous chloride load during resuscita-
N Engl J Med 2012;367:124-34. tion and in-hospital mortality among patients with SIRS.
Potura E, Lindner G, Biesenbach P, et al. An acetate-buffered Intensive Care Med 2014;40:1897-905.
balanced crystalloid versus 0.9% saline in patients with

CCSAP 2018 Book 3 • Fluids and Nutrition/GI and Liver Disorders 18 Fluid and Hyponatremia Management
Smith CA, Duby JJ, Utter GH, et al. Cost-minimization management in patients undergoing abdominal aor-
analysis of two fluid products for resuscitation of criti- tic aneurysm repair: an outcome study. Anesth Analg
cally injured trauma patients. Am J Health Syst Pharm 2001;93:817-22.
2014;71:470-5.
Young P, Bailey M, Beasley R, et al. Effect of a buffered crys-
Sood L, Sterns RH, Hix JK, et al. Hypertonic saline and des- talloid solution vs saline on acute kidney injury among
mopressin: a simple strategy for safe correction of severe patients in the intensive care unit: the SPLIT randomized
hyponatremia. Am J Kidney Dis 2013;61:571-8. clinical trial. JAMA 2015;314:1701-10.

Verbalis JG, Goldsmith SR, Greenberg A, et al. Diagnosis, Yunos NM, Bellomo R, Hegarty C, et al. Association between
evaluation, and treatment of hyponatremia: expert panel a chloride-liberal vs chloride-restrictive intravenous fluid
recommendations. Am J Med 2013;126:S1-S42. administration strategy and kidney injury in critically ill
adults. JAMA 2012;308:1566-72.
Waters JH, Gottlieb A, Schoenwald P, et al. Normal saline
versus lactated Ringer’s solution for intraoperative fluid

CCSAP 2018 Book 3 • Fluids and Nutrition/GI and Liver Disorders 19 Fluid and Hyponatremia Management
Self-Assessment Questions
1. The ICU attending calls for a recommendation on intra- 4. Which one of the following fluid therapy strategies would
venous fluid (IVF) administration for a patient who was be best to recommend for A.Y.?
admitted for a chronic obstructive pulmonary disease
A. 500 mL of 5% albumin over 15 minutes
exacerbation and is now hypotensive. The patient’s labo-
B. 0.9% sodium chloride 200 mL/hour continuous
ratory values are as follows: Na 142 mEq/L, K 3.9 mEq/L,
infusion
Cl 102 mEq/L, CO2 24 mEq/L, BUN 8 mg/dL, and glucose
C. Lactated Ringer solution 100 mL over 5 minutes
110 mg/dL. The patient has no history of heart failure
D. Furosemide 40 mg intravenously every 8 hours
and has a recent left ventricular ejection fraction of 60%.
Which one of the following is the best IVF administration 5. Two days later, A.Y.’s vasopressors are discontinued, and
to recommend for this patient? she remains hemodynamically stable with mean arte-
rial pressures of 70–80 mm Hg. She remains intubated
A. Hydroxyethyl starch 50-mL/kg bolus
and sedated on fentanyl 100 mcg/hour and propofol 20
B. 5% albumin 250-mL bolus
mg/kg/minute. Her blood and peritoneal fluid cultures
C. 0.9% sodium chloride 1000-mL bolus
have grown Enterobacter cloacae, and her antibiotics
D. 0.9% sodium chloride 100 mL/hour
were changed to ertapenem 1 g intravenously daily. Her
2. A 29-year-old man is in the ICU for sepsis secondary to abdominal drain has put out around 200 mL of fluid per
necrotizing fasciitis. After receiving fluid resuscitation of day. Currently, A.Y.’s laboratory findings include Na 144
30 mL/kg of normal saline, his Na is now 152 mEq/L and mEq/L, K 4.4 mEq/L, Cl 104 mEq/L, CO2 24 mEq/L, BUN
Cl is 116 mEq/L. The patient remains hypotensive and 33 mg/dL, SCr 2.1 mg/dL, and glucose 144 mg/dL. Her
has been determined to be fluid responsive. Which one of chest radiography reveals no significant changes, and
the following IVFs is best to recommend for this patient? her lungs are clear to auscultation. Further assessment
reveals that A.Y. is fluid responsive. Which one of the fol-
A. 0.9% sodium chloride
lowing is the best fluid management strategy for A.Y.?
B. 25% albumin
C. 5% albumin A. Furosemide 10-mg/hour continuous infusion
D. Plasma-Lyte 148 B. 0.9% sodium chloride 1500 mL over 60 minutes
C. Lactated Ringer solution 100-mL/hour continuous
Questions 3–5 pertain to the following case. infusion
A.Y., a 48-year-old woman (weight 53.5 kg), was admitted for D. 100 mL of 25% albumin over 30 minutes
an emergency bowel resection, in which she had an explor- 6. A 71-year-old man, who was admitted after falling from
atory laparotomy. Postoperatively, she is in the ICU with a ladder, is now in the ICU with several rib fractures
septic shock and respiratory failure. A.Y. has received 4 L of and a right femur fracture. His blood pressure is 73/44
0.9% sodium chloride over the past 3 hours and is currently mm Hg. On physical examination he appears to be fluid
intubated, on mechanical ventilation, and receiving fentanyl responsive. The patient currently has one peripheral
150 mcg/hour and midazolam 2 mg/hour for analgesia and intravenous catheter available for access. He is receiv-
sedation. A.Y. is also receiving piperacillin/tazobactam 4.5 g ing an infusion of packed RBCs for suspected bleeding
intravenously every 6 hours for intra-abdominal sepsis. She with an Hgb of 5.3 g/dL. Other laboratory values include
is receiving norepinephrine 30 mcg/minute and vasopressin Na 143 mEq/L, K 4.4 mEq/L, Cl 105 mEq/L, CO2 21 mEq/L,
0.04 unit/minute for shock, and her mean arterial pressure BUN 13 mg/dL, and SCr 1.1 mg/dL. Which one of the fol-
is 50 mm Hg. Laboratory values include Na 141 mEq/L, K 4.2 lowing fluids is best to recommend for this patient?
mEq/L, Cl 104 mEq/L, CO2 18 mEq/L, BUN 12 mg/dl, SCr 1.1
A. 5% albumin
mg/dL, and glucose 129 mg/dL.
B. 3% sodium chloride
3. Which one of the following would best assess A.Y.’s fluid C. 0.9% sodium chloride
responsiveness? D. Lactated Ringer solution
A. Ultrasound monitoring of the inferior vena cava (IVC) 7. An ICU patient with ascending cholangitis has received 10
B. Lactate measurement L of 0.9% sodium chloride and 2 L of lactated Ringer solu-
C. Esophageal Doppler monitoring tion for fluid resuscitation. She remains on vasopressor
D. Central venous pressure support and, despite the aggressive fluid administration,

CCSAP 2018 Book 3 • Fluids and Nutrition/GI and Liver Disorders 20 Fluid and Hyponatremia Management
appears to be intravascularly depleted. She also has had C. Normosol 1000-mL bolus
increasing oxygen requirements on the ventilator with D. 0.9% sodium chloride bolus
worsening pulmonary edema on chest radiography. Her
11. For which one of the following patients would peripheral
Na and Cl are 143 mEq/L and 102 mEq/L, respectively.
administration of 3% sodium chloride be most appropri-
Which one of the following is best to recommend for this
ate to treat hyponatremia?
patient?
A. 53-year-old man with chronic alcohol use disorder
A. 5% albumin
who presents with Na 125 mEq/L and is awake, alert,
B. 0.9% sodium chloride
and oriented
C. 3% sodium chloride
B. 40-year-old woman with 22-gauge peripheral
D. Plasma-Lyte 148
intravenous catheter and Na 118 mEq/L
8. A 36-year-old woman with a history of type 2 diabetes C. 66-year-old man with altered mental status, seizure,
is in the ICU for diabetic ketoacidosis. She has already Na 112 mEq/L, and significant difficulty inserting
received 8 L of 0.9% sodium chloride and requires addi- central venous access
tional volume. Her Na is 143 mEq/L, K 3.8 mEq/L, Cl 120 D. 23-year-old woman who presents with venlafaxine-
mEq/L, HCO3 17 mEq/L, and pH 7.19. Which one of the induced syndrome of inappropriate antidiuretic
following is best to recommend for additional fluid resus- hormone (SIADH) with Na 118 mEq/L
citation in this patient?
Questions 12 and 13 pertain to the following case.
A. 0.9% sodium chloride
B. Sodium bicarbonate L.T. is a 68-year-old man (weight 62 kg) with a history of
C. Plasma-Lyte 148 hypertension. He presents to the ED with severe altered
D. 25% albumin mental status. Medications before admission include hydro-
chlorothiazide 25 mg orally daily, lisinopril 10 mg orally daily,
9. A 36-year-old man was admitted 3 days ago after devel-
and spironolactone 100 mg orally daily. L.T.’s laboratory find-
oping respiratory failure from a heroin overdose. He
ings are Na 102 mEq/L, K 4.9 mEq/L, Cl 110 mEq/L, CO2 25
was hypotensive on admission and received 6 L of lac-
mEq/L, BUN 27 mg/dL, SCr 1.4 mg/dL, and glucose 110 mg/
tated Ringer solution for resuscitation. He has since
dL. On physical examination, no edema is noted; the patient
been hemodynamically stable and is being assessed for
has dry mucous membranes and poor skin turgor. His blood
readiness to extubate. Chest radiography reveals mild
pressure is 110/66 mm Hg, heart rate 120 beats/minute,
pulmonary edema. Which one of the following is best to
respiratory rate 24 breaths/minute, and SaO2 95%. L.T.’s wife
recommend for this patient?
reports he was evaluated by his primary physician about a
A. Administer an additional 1-L bolus of 0.9% sodium week ago, who increased his diuretic doses for poorly con-
chloride. trolled hypertension.
B. Start 0.45% sodium chloride/5% dextrose in water
12. Which one of the following is the optimal recommended
solution at 80 mL/hour for maintenance.
rate of change (increase) in L.T.’s sodium?
C. Administer 500 mL of 5% albumin solution.
D. Administer 20 mg of intravenous furosemide. A. 8 mEq/L in 24 hours
B. 10 mEq/L in 24 hours
10. A 50-year-old man is in the ICU with septic shock; his mul-
C. 12 mEq/L in 24 hours
tidisciplinary care team is discussing the plan for fluid
D. 18 mEq/L in 48 hours
therapy. The patient is intubated, sedated on fentanyl 150
mcg/hour and propofol 20 mcg/kg/minute, and receiving 13. L.T. was initially treated with 3% sodium chloride 100
norepinephrine 20 mcg/minute for shock. Currently, his mL/hour with a subsequent increase in Na to 115 mEq/L
mean arterial pressure is 68 mm Hg. On examination, the after 16 hours of receiving the infusion. Which one of
patient is not fluid responsive. He has received 12 L of the following strategies is best to recommend for L.T.’s
normal saline during his ICU stay, and chest radiography hyponatremia?
from this morning reveals significant pulmonary edema. A. Discontinue 3% sodium chloride; administer
Laboratory analysis shows Na 142 mEq/L, K 4 mEq/L, Cl conivaptan 20 mg intravenously × 1.
111 mEq/L, CO2 18 mEq/L, BUN 21 mg/dL, SCr 1.4 mg/dL, B. Discontinue 3% sodium chloride; administer
and albumin 2.4 g/dL. Which one of the following is the desmopressin 2 mcg intravenously every 8 hours
best approach for this patient’s fluid therapy? plus 5% dextrose in water 180 mL/hour.
A. 25% albumin 50-mL bolus C. Continue 3% sodium chloride at 100 mL/hour.
B. 0.9% sodium chloride 75 mL/hour D. Continue 3% sodium chloride at 100 mL/hour; add
desmopressin 2 mcg intravenously every 8 hours.

CCSAP 2018 Book 3 • Fluids and Nutrition/GI and Liver Disorders 21 Fluid and Hyponatremia Management
14. A 71-year-old man with a history of hyperlipidemia pres- 15. The ICU attending calls you for a recommendation on a
ents to the ED with 2-day history of dizziness and severe patient admitted with acute hyponatremia. She explains
diarrhea. On admission, his laboratory findings include that the patient has a peripheral line inserted with Na of
Na 122 mEq/L, K 3.1 mEq/L, Cl 94 mEq/L, CO2 18 mEq/L, 115 mEq/L and has significant peripheral edema, so she
BUN 33 mg/dL, and SCr 1.5 mg/dL. The patient reports would like to limit volume administration. Which one of
that he has not felt well for several days and has had lim- the following is best to recommend for this patient?
ited oral intake with the onset of diarrhea. On physical
A. Administer 3% sodium chloride 40 mL/hour for
examination, he has dry mucous membranes and poor
48 hours.
skin turgor. The patient was initiated on 0.9% sodium
B. Administer a 20-mg intravenous bolus of conivaptan
chloride 100 mL/hour through a peripheral intravenous
and reassess sodium in 4 hours.
line and desmopressin 2 mcg intravenously every 8
C. Administer a 20-mg intravenous bolus of conivaptan
hours on admission. Twenty-four hours later, his Na is
followed by a 20-mg intravenous continuous
129 mEq/L. Which one of the following is best to recom-
infusion for 5 days.
mend for this patient’s hyponatremia?
D. Administer 0.9% sodium chloride 100 m/hour for
A. Discontinue 0.9% sodium chloride, start 3% sodium 24 hours.
chloride 30 mL/hour, and continue desmopressin.
B. Continue 0.9% sodium chloride 100 mL/hour,
discontinue desmopressin, and administer
conivaptan 20 mg intravenously × 1.
C. Discontinue 0.9% sodium chloride, and continue
desmopressin.
D. Continue 0.9% sodium chloride 100 mL/hour, and
discontinue desmopressin.

CCSAP 2018 Book 3 • Fluids and Nutrition/GI and Liver Disorders 22 Fluid and Hyponatremia Management