Fluid Management in Acute Respiratory Failure

Introduction

Acute respiratory failure (ARF) is a challenging condition with significant variability in definition,¹ which encompasses patients with progressive and acute hypoxemia, hypercapnia or respiratory distress related to a variety of cardiorespiratory or systemic diseases such as sepsis. Acute Respiratory Distress Syndrome (ARDS) is a serious form of ARF defined by non-cardiogenic pulmonary edema and acute hypoxemia, triggered by various pulmonary and non-pulmonary insults.²

The ideal fluid management in ARF would avoid impairing respiratory mechanics and gas exchange while still ensuring adequate tissue perfusion. Critically ill patients (with hemodynamic instability) often require early intravenous fluid administration to support cardiac output, and inadequate initial fluid resuscitation may worsen mortality.³ While there are no standardized society guideline recommendations for optimal fluid management in ARF, the majority of our understanding of fluid management comes from research in patients with ARDS and sepsis.

This review will describe the pathophysiology of ARDS and how it relates to fluid management, considerations in patients with or at risk for ARDS, ARF and sepsis, fluid choice, post-hospital outcomes related to fluid management, metrics to guide fluid therapy, and provide a framework for fluid management depending on the clinical setting. A conservative or restrictive fluid strategy focuses on minimizing fluids during and after the resuscitative phase, and potentially adding active diuresis to reduce total fluid balance during non-resuscitative phase. A liberal fluid strategy has no restriction of intravenous fluids in the resuscitative phase and lacks protocolized active fluid removal during the recovery phase.

Considering the pathophysiology in ARDS in fluid management

In ARDS, healthy pulmonary endothelium is compromised with excessive inflammation leading to increased pulmonary endothelial and epithelial permeability, leakage of protein-rich fluid into the alveolar space, impaired fluid removal and disrupted alveolar barrier function.⁴ This can increase pulmonary hydrostatic pressures, decrease oncotic pressures and disrupt the pressure gradient leading to significant interstitial and alveolar flooding (i.e. pulmonary edema).⁴ Hypothesized benefits of a conservative strategy include decreasing hydrostatic pressures and right ventricular pressures, increasing oncotic pressures, and decreasing plasma levels of angiopoietin-2 which could have a protective effect on the vascular endothelium.⁵

Excessive fluid administration could exacerbate the underlying pathophysiology in ARDS and worsen pulmonary edema. However, hypotension and subsequent decreased cardiac output and pulmonary blood flow in patients with ARDS can worsen alveolar and physiologic dead space.⁶ Since the majority of patient with ARDS die from multiorgan system failure,⁷ resuscitation can be imperative. Unfortunately, the right strategy to maintain tissue perfusion while limiting pulmonary edema is uncertain.

A historical shift in fluid management

Researchers have compared liberal vs conservative strategies in patients with ARDS for more than 30 years.^8–10 Arguments for a liberal strategy, the historically conventional practice, were based on the benefit of end organ perfusion and oxygen delivery, hypothesizing that focusing on lowering the pulmonary capillary wedge pressures (PCWP) during pulmonary edema would not be beneficial and could be harmful in patients with multiorgan system involvement particularly in shock and acute kidney injury.¹¹ Additionally, there was significant provider practice variation in the dose of fluids given in early liberal strategies ranging from a fluid balance of 4 liters at day 4 in ARMA and ALVEOLI^12–13 and up to 20L at day 14 in an RCT of 113 ARDS patients.¹⁴

Early studies suggested that a conservative strategy (guided by extravascular lung water assessments)¹⁵ in patients with ARDS and acute lung injury (ALI) may decrease pulmonary artery pressures¹⁵ and increase intensive care unit (ICU) and ventilator free days and that a negative cumulative fluid balance and reduced PCWP could improve overall survival.^14,16–18Although findings were limited by the observational nature of some studies and significant confounding by severity of illness and small cohort size, it signified a paradigm shift.

This debate persisted until the landmark large multicenter randomized controlled trial (RCT), the Fluid and the Catheter Treatment Trial (FACTT), was published in 2006. Using a factorial design, FACTT included 1001 patients within 48 hours of diagnosis of ALI or ARDS randomized to either a conservative or liberal fluid strategy, guided by either pulmonary artery catheters (PAC) or central venous catheters (CVC).¹⁹ FACTT used strict protocols to guide active diuresis, vasoactive medications and fluid boluses and targeted a higher central venous pressure (CVP) or PCWP in the liberal group and a lower goal in the conservative group. The protocol deferred to clinician-driven management in the presence of shock and held diuresis for 12 hours. While there was no difference in the primary outcome of 60-day mortality (25% vs. 28%, P=0.30), the conservative group had significantly more ventilator and ICU-free days and a significant difference in cumulative fluid balance (−136 ml in the conservative group vs. nearly 7L in the liberal group). Notably, a conservative approach did not increase the incidence or duration of shock or acute kidney injury. Also importantly, there was no benefit of randomization to pulmonary artery catheter-based guidance. These studies suggest that there are potential clinical benefits and notably minimal harm with a conservative approach to fluid management in patients in ARDS who have no evidence of ongoing shock. Further studies are needed to understand the effect of fluid management strategies on mortality, management in patients with ARDS and concomitant shock, and the optimal approach to fluid and diuretic administration.

Fluid management in patients at risk of ARDS

Research in strategies to prevent development and progression of ARDS is lacking and findings in published observational studies are limited by potential indication bias. However, a positive fluid balance in a patient at risk of ARDS has been associated with a higher incidence of ARDS and increased mortality.²⁰An analysis of 455 patients with sepsis identified that 92% of patients who developed ARDS had hypoproteinemia at study entry and the severity of hypoproteinemia was identified as the strongest predictor for development of ARDS²¹, a finding confirmed in patients with septic shock.²² The low oncotic pressure associated with hypoproteinemia can facilitate hydrostatic pulmonary edema.

It is unclear how to mitigate the risk of hypoproteinemia. Colloids (such as albumin) may increase oncotic pressure and minimize volume overload. Although use of colloids has been evaluated in ARDS and acute lung injury (ALI),^23–25 no studies exist in patients at risk of ARDS. Use of other fluid types (i.e. hypertonic saline compared to normal saline) are also mixed.^26,27 A recent large multicenter RCT, Crystalloid Liberal or Vasopressors Early Resuscitation in Sepsis (CLOVERS) trial, compared a restrictive strategy focusing on early vasopressor use to a liberal strategy during the first 24 hours of treatment in 1,563 patients with sepsis-induced hypotension and did not demonstrate a difference between the groups in the development of ARDS within the first 7 days of randomization.²⁸ More research is required on the impact of fluid management in this population.

Fluid management in ARF and septic shock

ARF rarely occurs in isolation, and is common in patients with sepsis.²⁹ Sepsis is a syndrome of dysregulated immune response to severe infection causing organ dysfunction with hypotension and hypoperfusion, associated with a short-term mortality of up to 15%.³⁰ Treatment of patients with sepsis includes prompt delivery of antibiotics for the infection and administration of intravenous fluids and vasopressors to address hypoperfusion and shock.^31,32 ARF and sepsis are linked by multiple causal pathways. A pulmonary infection can cause respiratory failure and sepsis, non-pulmonary infections with sepsis are a risk factor for ARDS,³³ and treatment for sepsis with intravenous fluids can lead to iatrogenic respiratory failure by pulmonary edema.³⁴ As a result, fluid management in patients with sepsis and ARF can present opposing priorities for the clinician; fluid resuscitation to support perfusion and fluid removal for respiratory failure. This dilemma creates two research questions: first, whether the fluid strategy affects outcomes for patients with sepsis and ARF, and second, whether fluid strategy for sepsis affects the risk of subsequent respiratory failure.

In resuscitation for sepsis, guidelines have recommended initial treatment with at least 30 milliliters /kilogram (ml/kg) of intravenous fluids based largely on a RCT in 2001 that administered this large, weight-based volume of fluid as a component of multiple interventions given early in the care of patients with sepsis, termed early goal-directed therapy (EGDT).³² Similarly, an observational study of patients with ARDS from sepsis suggested that inadequate initial resuscitation during usual care was associated with worse survival.²⁹ Three subsequent multi-center trials (ProCESS³⁵, ARISE³⁶, ProMISe³⁷) compared EGDT to usual care for sepsis. A patient-level meta-analysis of these trials found that the EGDT strategy did not affect mortality overall or in the subgroup of patients with ARF on invasive mechanical ventilation (Adjusted Odds Ratio (AOR) 0.87 (0.55–1.37), p=0.55).³⁸ For respiratory failure as an outcome of fluid strategy, EGDT was associated with a small increase in the odds of receiving mechanical ventilation (AOR 1.05 (0.89–1.24), p=0.04). These studies demonstrated that EGDT protocols did not provide a benefit over usual care, but the degree to which additional fluid in the EGDT group affected respiratory failure is complicated by the comparison of the multiple EGDT therapies to variable usual care, which still included substantial fluid use.

In resource-limited settings, multiple trials have shown more direct evidence of harm from fluid administration in sepsis, particularly related to respiratory failure. In Zambia, the Simplified Severe Sepsis Protocol (SSSP-1) tested a sepsis protocol of early fluid use among 109 patients with severe sepsis and the trial was stopped early due to higher mortality in patients with hypoxemia randomized to the protocol.³⁹ The subsequent Simplified Severe Sepsis Protocol −2 (SSSP-2) trial excluded patients with respiratory failure and stopped giving fluids if patients were developing respiratory failure.⁴⁰ The group receiving the sepsis protocol received more fluids and had a higher rate of in-hospital mortality (48.1% vs 33.0%, p=0.03). More patients in the sepsis protocol group experienced worsening hypoxemia or tachypnea as well (36% vs 22% (p=0.03)).

In high-resource settings, two prospective trials inform fluid administration for patients with sepsis and ARF. The Fluid Responsiveness Evaluation in Sepsis-Associated Hypotension (FRESH) was a multicenter trial that randomized 150 patients with septic shock to an intervention arm with protocolized testing of fluid responsiveness to guide fluid therapy for the first 72 hours vs usual care. The protocolized group received a lower total volume of fluid compared to usual care with an average of 1.4L less cumulative fluid balance, the study’s primary outcome. Fewer patients randomized to the intervention arm subsequently received mechanical ventilation (17.7% vs 34.1%, p=0.04), a provocative finding suggesting need for further evaluation.⁴¹

The CLOVERS trial more specifically addressed the effect of fluid volume during resuscitation for sepsis on patient outcomes, including respiratory sequelae.¹⁶ Overall, 9.4% of patients had respiratory failure requiring assisted ventilation. In the first 24 hours after enrollment, patients in the restrictive strategy group received a median 1.3L of fluid (IQR 0.6–2.3) and those in the liberal group received 3.4L (IQR 2.5–4.5L). Both groups experienced a similar primary outcome of 90-day mortality, and all respiratory outcomes were also similar. While CLOVERS demonstrated the safety of a restrictive approach to initial fluid resuscitation in sepsis, it did not suggest a unique benefit for respiratory failure. Two ongoing additional trials of initial fluid-volume strategy in sepsis (NCT04569942, NCT05179499) will further inform this question.

After an initial bolus of fluid for patients with sepsis, additional fluids to maintain hemodynamic goals may be beneficial, ineffective, or harmful. Several observational studies have found a more positive fluid balance is associated with increased morbidity and mortality.^34,42The recent Conservative versus Liberal Approach to Fluid Therapy of Septic Shock in Intensive Care (CLASSIC) trial compared fluid management strategies after initial resuscitation among 1,554 critically ill patients with sepsis and within 12 hours of the onset of shock.⁴³ The restrictive group was permitted fluid only under strict physiologic criteria to treat severe hypoperfusion with small volume boluses, while the standard-fluid group was not limited. At 5 days after enrollment, the median difference in fluid administered was 1.9L and in cumulative fluid balance between groups was 0.7L. The primary outcomes of death at 90 days showed no difference between the groups. Assessing the sub-group of patients with respiratory failure on ventilatory support at enrollment, however, demonstrated heterogeneity of treatment effect (p=0.03) with an absolute difference in mortality of −5.1% (95%CI: −11.3%–1.6%) favoring the restrictive strategy. Among all patients in the trial, both groups had a similar respiratory failure outcome in ventilator-free days. Despite the relatively small separation between groups achieved by the restrictive fluid strategy, these data provide provocative evidence supporting a restrictive strategy for fluid administration in patients with sepsis and respiratory failure after the initial fluid bolus.

After resolution of shock from sepsis, very few trials have prospectively studied strategies for fluid removal by diuresis. In the FACTT trial, 70% of patients had sepsis or pneumonia, suggesting that conclusions about the benefit of fluid limitation and diuresis in ARF can also be applied for patients with sepsis and ARF. A feasibility trial comparing a conservative fluid management strategy including diuretics to usual care enrolled 30 patients recovering from sepsis and was unable to achieve a meaningful difference in mean daily fluid balance to continue with a larger clinical trial, highlighting the difficulties in achieving substantial fluid removal in critically ill adults.⁴⁴

Until recently, few trials had addressed fluid management strategies for patients who had both ARDS and sepsis.⁴⁵ While patients with sepsis are felt to need some intravenous fluids for resuscitation to meet physiologic targets, the optimal amount is not known. But recent data, particularly from CLOVERS and CLASSIC trials, suggest that a more conservative fluid management strategy in patients with respiratory failure and sepsis are safe and that it may even be beneficial in patients in the days following resuscitation. Across many trials, patients with sepsis are at risk of respiratory failure and protocols that increase fluid administration increase this risk. Each trial protocol provides different measures by which to individualize the treatment of sepsis with fluids to achieve physiologic parameters (e.g. monitoring strategies, measures of fluid responsiveness). Minimizing variation in protocols is an important area for future trials to elucidate treatment effects and identify the best fluid management strategy for most patients.

Long-term outcomes of fluid management in ARDS

There is a lack of studies informing the impact of fluid management strategy on long-term outcomes after ARF. Jolley and colleagues conducted a secondary analysis of the FACTT cohort to determine whether race modified the effect of fluid administration strategy on 1-year morality after ARDS.⁴⁹The study included 655 non-Hispanic white and Black patients and demonstrated that conservative fluid management was associated with improved 1 year survival after ARDS only for Black patients (38% vs. 54%, P= 0.03) and overall, there was a significant interaction between identified race and fluid treatment group (P=0.012). While more studies are needed in this area, this suggests that the effect of fluid management strategy on longer term outcomes may not be homogenous across patient subgroups.

Mikkelsen and colleagues conducted neuropsychological testing of survivors enrolled in the FACTT trial by phone at 2 and 12 months after hospital discharge, and 75 patients (7.5% of the patients in FACTT) completed all testing.^46,47 Randomization to the conservative fluid management arm was independently associated with cognitive impairment at follow-up. This surprising finding has no obvious mechanistic cause, as physiologic parameters like cardiac index or systolic blood pressure were similar between the study groups and thus do not suggest reduced cerebral perfusion occurred in the conservative fluid group. The sample of patients in this secondary analysis was highly vulnerable to selection bias due to patients lost to follow up or the competing risk of death, which limits the interpretation of this study.

Among patients with septic shock who were enrolled in the CLASSIC trial, a follow up study obtained 1-year accounted for cognitive outcomes of 86.3% of enrolled patients Mini MoCA tool⁴⁸ in 568 survivors. However, there was no difference in cognitive testing among survivors (adjusted odds ratio 1.01 (95%CI: 0.95–1.07)) or all patients accounting for those who were known to have died. The authors also found no difference in mortality or health-related quality of life by the fluid strategy assigned in the trial. Overall, the potential impact of fluid management strategies on post-hospital patient-centered outcomes in survivors of ARF remains largely unknown.

Fluid choice in ARF and ARDS

Ideally, intravenous fluids would provide hemodynamic benefit by expanding intravascular volume to increase cardiac output without worsening interstitial edema like that seen in ARDS. No fluid type is without physiologic complications, but those available for use can be divided into two categories: colloids and crystalloids. Colloids such as human serum albumin are fluids that contain large molecules such which cannot readily pass through healthy capillary membranes into interstitial spaces, while crystalloids are solutions of electrolytes in water that can easily do so.

Colloids are of particular interest in patients with respiratory failure given hypoproteinemia is associated with ARDS as a risk factor. The Saline vs. Albumin Fluid Evaluation (SAFE) trial randomized 6,997 critically ill adult patients to a 4% albumin solution or normal saline for resuscitation. There was no difference in mortality among the groups and among the prespecified subgroup of patients with ARDS.⁵⁰ A follow up trial, the Albumin Italian Outcome Sepsis (ALBIOS) trial, enrolled only patients with sepsis and used 20% albumin to achieve a target serum album level.⁵¹The ALBIOS trial found no difference between groups in the primary outcome, nor were the secondary and tertiary respiratory outcomes different.

In addition to its role as a resuscitation fluid, colloids have also been used to facilitate fluid removal. Martin and colleagues randomized 37 patients with hypoproteinemia and ARDS to receive furosemide and albumin compared to furosemide with placebo.²⁴ The combination of albumin and furosemide improved outcomes, demonstrating the albumin was beneficial for diuresis but showed no difference in patient-centered outcomes.²⁴A systematic review and meta-analysis of the 206 patients randomized to colloid therapy vs crystalloid in ARDS reinforced the limitations of these available data and the importance of larger trials to inform this question.⁵²

Crystalloid fluids are one of the most commonly administered therapies to hospitalized adults, and they are typically categorized into unbalanced (e.g. saline (0.9% sodium chloride)) or balanced (e.g., Lactated Ringer’s, Normosol-R) solutions. While saline includes supra-physiologic concentrations of chloride, balanced solutions contain an organic anion which allows for a reduced concentration of chloride and a more neutral pH compared to the more acidemic saline. Several large, randomized trials have been completed in the last 10 years to compare balanced solutions vs saline^53–56and a Bayesian meta-analysis using data from 34,450 trial participants concluded with an 89.5% probability that balanced crystalloid solutions were associated with lower mortality compared to saline.⁵⁷ No analysis examined the subgroup with respiratory failure, but these patients were well represented as many of the thousands of critically-ill patients were receiving invasive mechanical ventilation in the SMART⁵³ (34%), BASICS⁵⁵ (44%) and PLUS⁵⁶(79%) trials. These trials found no effect from crystalloid fluid types on the number of days alive and free of mechanical ventilation outcome, suggesting that patients with respiratory failure are unlikely to derive unique benefit from balanced solutions.

Review of proposed metrics to guide fluid resuscitation in ARF

Diagnostic metrics ideally should be used to assess hemodynamics and guide fluid responsiveness in ARF particularly in patients with persistent circulatory failure (shock) where the goal is to resuscitate and optimize tissue perfusion. An ideal metric would be easy and safe to measure, consistent, reliable in predicting fluid responsiveness, and proven to lead to improved patient outcomes. No such metric has been identified solely for this purpose.

Extravascular lung water (EVLW)

EVLW is the volume of water in the lungs (normal index is < 7ml/kg of ideal body weight) and a prognostic predictor of mortality in critically ill patients and ARDS.^58–60 EVLW can correlate with the severity of pulmonary edema with an increase up to 40–50 ml/kg body weight and it has been recommended as a potential quantitative parameter to guide fluid management in ARDS. An RCT of 101 critically ill patients compared a protocol of fluid management targeting a lower EVLW to a protocol based on PCWP with a primary outcome of development or resolution of pulmonary edema ¹⁵ Fluid restriction based on EVLW was associated with fewer ventilator, ICU days and a nearly neutral cumulative fluid balance over 3 days compared to a 2 liter fluid excess in the PCWP group. Another small study comparing EVLW to PCWP for fluid management in 29 patients with ARDS found that utilizing EVLW reduced ventilator and ICU days.⁶¹ However, while EVLW may be a useful tool to indicate fluid accumulation in lungs, the diagnostic framework for bedside measurement (transpulmonary thermodilution) and interpretation can be complex and it is not widely utilized in clinical practice yet. Further larger studies are needed and recently completed (HEAL Study, NCT00624650) to confirm its use for lung targeted fluid strategies compared to metrics such as CVP and urine output in management of ARF.

CVP and PAWP

CVP, commonly measured in critically ill patients via an indwelling CVC, may be used to measure intravascular volume as a surrogate for volume overload or RV failure. Studies have demonstrated poor correlation with other measures of fluid status. However, the FACTT study demonstrated that a CVP-guided strategy decreased duration of mechanical ventilation.¹⁹ Notably, in the last 20 years since, fewer patients are managed with CVCs, further challenging implementation of the FACTT approach. Reducing PAWP in ARDS with fluid restriction had been suggested to reduce mortality and increase ventilator free days although studies have been limited by exclusion criteria and difficulty generalizing results to all ARDS patients¹⁵and sample size.⁶⁰ Overall, the routine use of the PACs to measure PAWP and guide fluid management is not recommended due to increased complications compared to CVCs with no improvement in mortality or organ function.⁶²

Pulse pressure variation (PPV)

The respiratory variation of continuous PPV – difference between systolic and diastolic pressure – may be utilized as a reflection of ventricular stroke volume and and position on the Frank-Starling curve in patients with sinus rhythm. However, this method has not been validated and underperforms in patients with ARDS particularly due to the use of low tidal volumes in this population, and it may lead to increased false positive rates in patients with right ventricular dysfunction (a marker of potential fluid overload).^{63, 64}

Echocardiography/ ultrasound

The use of lung ultrasound to measure EVLW and detect aeration changes related to fluid administration may be used to detect pulmonary edema in patients with ARDS.^65,66 However, studies were small and the role of this tool in guiding fluid management in ARF needs further clarification. Critical care echocardiography is another important tool to assess hemodynamics in ventilated patients. Dynamic parameters such as the respiratory variations of the inferior vena cava diameter and doppler velocity in the left ventricular outflow tract may be used to predict fluid responsiveness in ventilated patients. The ratio of right ventricle to left ventricule end-diastolic area could also provide evidence of RV dilatation and RV failure, with some limitations such as inter-operator variability.^67,68 End-expiratory and end-inspiratory occlusion changes on aortic doppler velocity can accurately predict fluid responsiveness in critical ill mechanically ventilated patients with ARF.⁶⁹ Other parameters such as measuring respiratory variation of the superior vena cava⁶⁷ may be useful in ARDS patients due to high sensitivity to predict fluid responsiveness. However, this is not a readily accessible metric as measurement requires the use of transesophageal echocardiography.

Specific considerations in ARF and related fluid strategy

Surgical patients

Patients with ARDS due to trauma may benefit from a strategy of diuresis and albumin to improve fluid balance and oxygenation.²³ Furthermore, a post-hoc analysis of surgical patients with ALI and ARDS in the FACTT cohort identified increased ventilator-free and ICU-free days with a conservative strategy.⁷⁰ Studies in the literature evaluating intra-operative and postoperative fluid administration in surgical patients have demonstrated an association between excessive fluid administration and incidence of postoperative pulmonary complications (ALI and ARDS) and a reduction in these complications in patients managed with a restrictive fluid strategy intraoperatively.⁷¹ Based on current evidence, a restrictive strategy may be beneficial in surgical patients to minimize risk of ARF however the ideal fixed volume of fluids and type of fluid therapy are not well defined.⁷²

Individualized approach

Given the multiple different etiologies of ARF with varying responses to treatment, this may lead to difficulty identifying a unifying strategy for fluid management. One proposal to address heterogeneity is identifying sub-phenotypes to distinguish who will respond to a specific therapy. Through latent class analysis of ARDS cohorts, two sub-phenotypes (hyperinflammatory and less inflammatory) were recently identified and have shown differing responses to ventilator strategies and inflammatory markers.⁷³ Famous and colleagues completed a secondary analysis of the FACTT study and demonstrated that a conservative fluid management strategy lead to worse survival in the hyper-inflammatory phenotype of ARDS compared to the less inflammatory phenotype which had improved mortality with conservative fluid management.⁷⁴

Discussion

The accumulation of studies suggests fluid management in ARF is not as simple as an isolated liberal or conservative strategy for all patients. (Table 1). These strategies are complementary and clinicians should personalize fluid management based on hemodynamic assessment, the clinical context and response to ongoing therapy. (Figure 1) Ideally, this would mean consistently prioritizing a conservative fluid strategy in the absence of shock to facilitate weaning from mechanical ventilation and minimizing pulmonary edema.

Table 1.

Summary of select key studies assessing fluid management strategies in adult patients with or at risk of acute respiratory failure

Topic	Study name, Author, Year	Study design and aim	Population	N	Key findings
Patients with ARDS or ALI	FACTT19, NHLBI ARDS Clinical Trials Network, 2006	RCT comparing a conservative and liberal strategy within 48 hours of ARDS/ ALI diagnosis using strict protocols of fluid boluses, vasoactive medications or diuresis guided by specific CVP or PCWP goals	ALI or ARDS	1001	No difference in 60-day mortality (primary outcome), incidence or prevalence of shock or use of dialysis; improved oxygenation index and increased vent-free and ICU-free days with conservative strategy; significant difference in cumulative fluid balance (-136 ml in the conservative group compared to nearly 7L in liberal group)
	Sakr et al20, 2005	Multicenter observational study to evaluate variables and ventilator strategy associated with mortality	ALI and ARDS	3147	Positive fluid balance was associated with increased mortality
Patients at risk of ARDS	Mangiarlardi et al21, 2000	Multicenter observational study to evaluate outcomes in pts with hypoproteinemia compared to patients with normal serum total protein levels	Severe sepsis	455	92% of patients who developed ARDS had borderline or overt hypoproteinemia; associated with fluid retention, weight gain, prolonged mechanical ventilation and mortality
	Chang et al22, 2014	Retrospective cohort study to examine associations between volume of IV fluids administered in resuscitative phase of sepsis and septic shock and development of ARDS	Severe sepsis and septic shock	296	No significant association between volume of IV fluids administered in the 1st 24 hours of hospitalization and development of ARDS; serum albumin and APACHE II score on admission were most informative for development of ARDS
Patients with severe sepsis or septic shock	CLOVERS28, PETAL network, 2023	Multicenter RCT compared a restrictive fluid strategy (with early vasopressor usage) to a liberal fluid strategy in sepsis-induced hypotension	Severe sepsis and septic shock	1503	No difference in 90-day mortality (primary outcome), vent-free days, ICU-free days, intubation from mechanical ventilation by day 27 or ARDS onset between day 1 and day 7
	FRESH41, Douglas et al, 2020	Multicenter RCT evaluating whether resuscitation (fluid bolus or increase in vasoactive medications) guided by dynamic assessment of FR (PLR) compared to usual care improved patient outcomes	Septic shock	150	Fluid balance at 72 hours or ICU discharge (primary outcome) was significantly lower in intervention arm compared to usual care (1.3L vs. 2.02L, p=0.021); fewer patients in intervention arm required RRT (5.1% vs. 17.5%, p=0.04) or mechanical ventilation (17.7% vs. 34.1%, p=0.04)
	CLASSIC45, Meyhoff et al, 2016	Multicenter RCT comparing a restrictive strategy guided by strict physiologic criteria to standard care	Septic shock	1554	No difference in 90-day mortality (primary outcome); similar respiratory failure outcomes in vent-free days; in sub-group of patients with respiratory failure on ventilatory support (invasive or NIPPV) at enrollment, p-value for heterogeneity was 0.03 with an absolute difference in mortality of −5.1% favoring the restrictive fluid strategy
Fluid choice in ARF or ARDS	SAFE52, Finfer et al, 2004	Multicenter RCT comparing FR with 4% albumin or normal saline	Critically ill patients	6997	No difference in 28-day mortality among groups (primary outcome) and no difference in mortality among pre-specified subgroup of patients with ARDS; no difference in vent-free days
	ALBIOS53, Caironi et al, 2004	Multicenter RCT comparing 20% albumin and crystalloid or crystalloid alone	Severe sepsis	1818	No difference in 28-day or 90-day mortality among groups (primary outcome) or vent-free days
	Martin et al23, 2002	Multicenter RCT comparing 5 days of 25 g of albumin q8 hours and continuous diuresis (furosemide) or dual placebo targeted to diuresis, weight loss, and serum total protein	Mechanically ventilated pts with ALI and hypoproteinemia	37	Improvements in the PaO2/ FiO2 ratio in the treatment group within 24 hours (from 171 to 236, p=0.02); no difference in mortality or respiratory mechanics
	Martin et al24, 2005	Multicenter RCT comparing furosemide with albumin or furosemide with placebo for 72 hours, titrated to fluid loss and normalization of serum total protein concentration	Mechanically ventilated patients with ALI/ARDS and hypoproteinemia	40	Albumin-treated patients had greater increases in oxygenation (primary outcome) from baseline (mean change in PaO2/FiO2: +43 vs. −24 mm Hg at 24 hrs and +49 vs. −13 mm Hg at day 3), serum total protein (1.5 vs. 0.5 g/dL at day 3), and net fluid loss (−5480 vs. −1490 mL at day 3) throughout the study period (all p < .05)

Definitions of abbreviations: ARDS= Acute Respiratory Distress Syndrome; ALI= Acute Lung Injury; RCT= randomized controlled trial; ARF= acute respiratory failure; NHLBI= National Heart, Lung, and Blood Institute; CVP= central venous pressure; PCWP= pulmonary capillary wedge pressure; PLR= passive leg raise; ICU= intensive care unit; IV= intravenous; APACHE II= acute physiology and chronic health evaluation; RRT= renal replacement therapy; NIPPV= non-invasive positive pressure ventilation; FR= fluid resuscitation; PaO₂ = partial pressure of oxygen; FiO₂= fraction of inspired oxygen

Figure 1:

Acute Respiratory Failure in a Critically Ill Paitient

Credit Line: Image created using Miro

Gaps in literature

Future studies should address how to balance fluid strategy in patients with concomitant shock and ARF, long term outcomes, the optimal choice and quantity of fluid, timing of initiation of diuretics and the proper hemodynamic metrics to guide fluid management.

Summary

Research assessing fluid management in ARF has largely focused on patients with ARDS where hypoxemia can be exacerbated by leaky membrane permeability and non-cardiogenic pulmonary edema. While the optimal approach is unclear, studies suggest a conservative fluid management in patients who are not in shock in ARF patients is a safe strategy. More work is needed to understand the ideal approach for individual patients.

Clinics Care Points:

• There is evidence that a conservative fluid strategy in patients with ARDS who are no longer in shock is minimally harmful and should be the standard approach in this population. While studies suggest this may also be beneficial in patients with ARF and sepsis, more research is needed in patients at risk of ARDS.

• While fluids are commonly used in critically ill patients, this can be harmful if administered inappropriately and it is imperative to consider the indication, dose, timing and safety profile as analogous to the prescription of any other drug in critical care.

• Fluid management may influence long-term outcomes and research on developing a personalized approach to fluid management and the appropriate volume and type of fluid to administer is urgently needed.

Key Points:

1. Optimal fluid management in patients with acute respiratory failure is challenging to standardize, given the heterogeneity of timing and the associated acute and chronic conditions.

2. Liberal and conservative fluid strategies are likely complimentary and ongoing assessment in the same patient is recommended to optimize tissue perfusion and avoid volume overload.

3. A conservative fluid strategy following initial resuscitation may be beneficial with minimal harm in patients with ARDS and patients with concomitant respiratory failure and sepsis.

4. The ideal type of fluid, volume and timing of diuresis during shock and after resolution of shock in ARF remains unclear.

5. Studies identifying sub-phenotypes of patients that will benefit from conservative or liberal fluid management are urgently needed.

Synopsis:

Fluid management in acute respiratory failure is an area of uncertainty requiring a delicate balance of resuscitation and fluid removal to manage hypoperfusion and avoidance of hypoxemia. Overall, a restrictive fluid strategy (minimizing fluid administration) and careful attention to overall fluid balance may be beneficial after initial resuscitation and does not have major side effects. Further studies are needed to improve our understanding of patients who will benefit from a restrictive or liberal fluid management strategy.