Residency · Residency · Internal Medicine

Fluid Resuscitation: Crystalloids, Colloids, and Albumin

Introduction

Intravenous fluid therapy is one of the most common interventions in hospital medicine, yet the choice of fluid type, volume, and rate remains a source of considerable debate. Recent landmark trials have clarified the superiority of balanced crystalloids over normal saline in many clinical contexts and defined the limited role of colloids. A nuanced understanding of fluid physiology and the evidence base is essential for internal medicine residents.

Fluid Physiology

• Total body water: approximately 60% of body weight (42 L in a 70 kg male)
• Distribution: intracellular (2/3) and extracellular (1/3); extracellular divided into interstitial (3/4) and intravascular (1/4)
• IV crystalloids distribute primarily into the extracellular space; approximately 25% remains intravascular after 1 hour
• The revised Starling model and the endothelial glycocalyx govern transvascular fluid movement; the glycocalyx is degraded by sepsis, surgery, and hypervolemia, increasing capillary leak
• Effective circulating volume, not total body water, determines organ perfusion

Crystalloids

Normal Saline (0.9% NaCl)

• Composition: Na 154 mEq/L, Cl 154 mEq/L; pH 5.5; osmolality 308 mOsm/L
• Supraphysiologic chloride load: leads to hyperchloremic non-anion gap metabolic acidosis
• High chloride causes renal afferent arteriolar vasoconstriction, reducing GFR and urine output
• Appropriate use: hypochloremic metabolic alkalosis, hyponatremia, hyperkalemia (no potassium content), and as a diluent for certain medications

Balanced Crystalloids

• More closely approximate plasma electrolyte composition
• Lactate in LR is rapidly metabolized by the liver; does NOT worsen lactic acidosis

Key Trials: Balanced vs. Normal Saline

• SMART trial (2018, single-center): balanced crystalloids reduced the composite of death, new renal replacement therapy, and persistent renal dysfunction vs. saline in critically ill patients (absolute risk reduction 1.1%)
• SALT-ED trial (2018): similar findings in non-critically ill ED patients
• BaSICS trial (2021, multi-center): no significant difference between balanced solutions and saline in ICU patients (mortality or AKI); however, trend favoring balanced solutions in sepsis subgroup
• PLUS trial (2022): no significant difference between Plasma-Lyte and saline in ICU patients
• Current consensus: balanced crystalloids are preferred as the default resuscitation fluid; normal saline reserved for specific indications (hyponatremia, metabolic alkalosis, hyperkalemia, brain injury)

Colloids

Albumin

• 5% albumin: iso-oncotic; expands intravascular volume approximately 1:1 (500 mL infused yields approximately 500 mL intravascular expansion)
• 25% albumin (hyperoncotic): draws fluid from interstitial space; 100 mL expands intravascular volume by approximately 400-500 mL
• SAFE trial (2004): 4% albumin equivalent to normal saline for resuscitation in general ICU patients; HARM in traumatic brain injury subgroup (increased mortality)
• ALBIOS trial (2014): albumin + crystalloid vs. crystalloid alone in sepsis; maintained higher MAP but no mortality difference; post-hoc subgroup analysis suggested benefit in septic shock
• Indications for albumin:
• Spontaneous bacterial peritonitis: albumin 1.5 g/kg day 1, 1 g/kg day 3 (reduces renal failure and mortality)
• Large-volume paracentesis: 6-8 g albumin per liter removed after > 5 L
• Hepatorenal syndrome: albumin with terlipressin or norepinephrine
• Septic shock refractory to crystalloids (reasonable adjunct)
• ARDS with hypoproteinemia: albumin + furosemide may improve fluid balance

Hydroxyethyl Starch (HES)

• SHOULD NOT BE USED for fluid resuscitation
• 6S trial and CHEST trial: HES increased mortality and acute kidney injury requiring renal replacement therapy in sepsis and critically ill patients
• FDA black box warning against use in critically ill patients
• Removed from formularies in many countries

Other Colloids

• Dextrans: rarely used; risk of anaphylaxis, coagulopathy, AKI
• Gelatin-based solutions (Gelofusine): used in some countries; associated with anaphylaxis and AKI; insufficient evidence for routine use

Assessing Fluid Responsiveness

• Only approximately 50% of hemodynamically unstable patients are fluid responsive
• Static markers (CVP, PCWP) are poor predictors of fluid responsiveness; CVP should NOT guide fluid administration
• Dynamic assessments:
• Passive leg raise (PLR): elevate legs to 45 degrees; if cardiac output or pulse pressure increases by >= 10%, the patient is fluid responsive; works in spontaneous breathing and arrhythmias
• Pulse pressure variation (PPV): > 13% variation during positive pressure ventilation predicts fluid responsiveness (requires sinus rhythm and controlled ventilation with Vt >= 8 mL/kg)
• Stroke volume variation (SVV): similar utility to PPV
• IVC ultrasound: IVC diameter < 2.1 cm with > 50% collapse suggests low CVP but does NOT reliably predict fluid responsiveness; best used to identify volume overload (plethoric IVC)
• Mini-fluid challenge: 100-200 mL crystalloid over 1 minute; assess cardiac output response

Volume of Resuscitation

• Initial sepsis resuscitation: 30 mL/kg crystalloid within the first 3 hours (SSC 2021 guideline); however, this is a weak recommendation and should be individualized
• CLOVERS trial (2023): restrictive fluid strategy (vasopressors first) vs. liberal fluids in sepsis-induced hypotension showed no difference in mortality; supports individualized approach
• CLASSIC trial (2022): restrictive IV fluids in ICU patients with septic shock did not differ from standard care in 90-day mortality
• Conservative fluid management after initial resuscitation reduces ventilator days and ICU stay (FACTT trial in ARDS)
• General principle: resuscitate early, de-escalate quickly; transition from "rescue" to "maintenance" to "de-resuscitation" phases

Special Populations

• Traumatic hemorrhage: damage control resuscitation with balanced blood products (1:1:1 ratio of PRBC:FFP:platelets); limit crystalloid; permissive hypotension (target SBP 80-90 mmHg) until surgical control
• Diabetic ketoacidosis: initial bolus of normal saline (or balanced crystalloid), then transition based on corrected sodium
• Burns: Parkland formula (4 mL/kg/% TBSA over 24 hours); titrate to urine output 0.5-1 mL/kg/hr
• Traumatic brain injury: avoid hypotonic solutions (LR is mildly hypotonic); normal saline or hypertonic saline preferred

Key Clinical Pearls

• Balanced crystalloids (LR or Plasma-Lyte) should be the default resuscitation fluid; normal saline is appropriate for specific indications (hyponatremia, hyperkalemia, brain injury)
• Hydroxyethyl starch is harmful and should never be used for resuscitation
• CVP is a poor predictor of fluid responsiveness; use dynamic assessments (passive leg raise, pulse pressure variation) instead
• After initial resuscitation, shift to a conservative fluid strategy; every liter of unnecessary fluid contributes to organ edema and worsened outcomes

References

1. Semler MW, Self WH, Wanderer JP, et al. Balanced Crystalloids versus Saline in Critically Ill Adults (SMART). N Engl J Med. 2018;378(9):829-839.
2. Finfer S, Bellomo R, Boyce N, et al. A Comparison of Albumin and Saline for Fluid Resuscitation in the ICU (SAFE). N Engl J Med. 2004;350(22):2247-2256.
3. National Heart, Lung, and Blood Institute. Early Restrictive or Liberal Fluid Management for Sepsis-Induced Hypotension (CLOVERS). N Engl J Med. 2023;388(6):499-510.
4. Myburgh JA, Finfer S, Bellomo R, et al. Hydroxyethyl Starch or Saline for Fluid Resuscitation in Intensive Care (CHEST). N Engl J Med. 2012;367(20):1901-1911.

For the intern

At 3am, if your patient is hypotensive or showing signs of shock, start with a 500 mL bolus of balanced crystalloid (Lactated Ringer's or Plasma-Lyte) over 15-30 minutes and reassess. Use normal saline only for specific indications: hyperkalemia, severe hyponatremia, or traumatic brain injury. Before giving more fluid, do a passive leg raise test with vital sign monitoring or bedside ultrasound to check if they're actually fluid-responsive—most hypotensive patients aren't. If they're not responding after 1-2 liters, call your senior and consider vasopressors. The discussion about CVP targets, albumin indications, and conservative maintenance strategies can wait until morning rounds when you can review trends and lactate clearance with your team.

Common pitfalls

• Using CVP to guide fluid therapy: CVP does not predict fluid responsiveness. A "low" CVP doesn't mean give more fluid, and a "high" CVP doesn't mean stop. Use dynamic assessments like passive leg raise instead.

• Giving liters of normal saline by default: This causes hyperchloremic acidosis and acute kidney injury. Switch your mental default to balanced crystalloids unless there's a specific reason for saline.

• Continuing aggressive fluids beyond initial resuscitation: After stabilization, excess fluid causes pulmonary edema, gut edema, and worse outcomes. Transition quickly to conservative maintenance and consider diuresis once stable.

• Reaching for albumin without an indication: Outside of SBP, large-volume paracentesis, and hepatorenal syndrome, albumin offers no mortality benefit and costs significantly more than crystalloids.

Make it stick

You're seeing Mrs. Davies, a 78-year-old admitted with pneumonia, now hypotensive to 85/45 mmHg and looking unwell. Your senior resident asks for your fluid plan. You recall the lecture: "balanced crystalloids are default." You order a 500 mL bolus of Lactated Ringer's. As the bag infuses, you perform a passive leg raise. Her blood pressure rises to 105/60 mmHg, heart rate drops from 110 to 95 bpm, and she looks more alert. You've just confirmed she's fluid-responsive and given the right fluid.

If you remember one thing

Default to balanced crystalloids (Lactated Ringer's or Plasma-Lyte) for fluid resuscitation, and always assess fluid responsiveness before giving more than an initial bolus.

Read this lecture as Markdown