Intravenous fluids — crystalloids and colloids pharmacology

Definition and Classification

Crystalloids are intravenous fluid solutions containing solutes with molecular weights less than 30 kDa, typically sodium chloride, glucose, or both, along with other electrolytes. The name derives from their ability to crystallise. Unlike colloids, crystalloid molecules freely cross capillary membranes and distribute throughout the extracellular fluid (ECF) compartment.

Crystalloids are classified by their tonicity relative to plasma (osmolality ~285-295 mOsm/kg):

Composition of Common Crystalloids

Fluid Distribution and Volume Kinetics

Compartment Distribution

After IV infusion, crystalloid distribution is governed by Starling forces and membrane permeability. Because sodium is the primary extracellular cation and the cell membrane largely excludes sodium, isotonic crystalloids expand the entire ECF, both intravascular and interstitial compartments.

$$\text{Fraction intravascular} = \frac{\text{Plasma volume}}{\text{ECF volume}} \approx \frac{3\,\text{L}}{14\,\text{L}} \approx 21\%$$

In practice:

• A 1 L infusion of isotonic crystalloid results in < 250 mL expansion of intravascular volume
• The remaining ≈ 750 mL distributes to the interstitial space
• After 20-30 minutes, an estimated 75-80% of an infused isotonic crystalloid has left the circulation

Glucose-containing solutions behave differently, once glucose is metabolised, the solution effectively becomes free water distributed to all body fluid compartments including the intracellular fluid (ICF). The half-time for glucose metabolism in healthy volunteers is approximately 15 minutes, though this is considerably longer during surgery and critical illness.

Hypotonic solutions equilibrate with ECF, making it hypotonic relative to ICF, driving osmotic water entry into cells, undesirable in the context of raised intracranial pressure or cerebral oedema.

Hypertonic Crystalloids

Hypertonic saline (3%, 7.5%) has osmolality far exceeding ECF. It rapidly mobilises water from the intracellular compartment to the ECF, producing a rapid but transient plasma volume expansion. If hypertonic solutions do not contain colloids, the plasma volume expansion is largely gone within 1-2 hours.

Mechanisms of Acid-Base Effects

Normal Saline and Hyperchloraemic Metabolic Acidosis

Normal saline is not "normal", its chloride concentration (154 mmol/L) far exceeds plasma (95-105 mmol/L). Large-volume administration causes:

1. Dilutional effect: bicarbonate buffering system diluted across expanded ECF
2. Strong Ion Difference (SID) effect: The SID (sum of strong cations minus strong anions) should approximate +40 mEq/L in plasma. Adding a solution with equal Na⁺ and Cl⁻ reduces SID, obligating a fall in pH per Stewart's physicochemical framework:

$$\text{SID} = [\text{Na}^+] + [\text{K}^+] + [\text{Mg}^{2+}] - [\text{Cl}^-] - [\text{Lactate}^-] \approx +40\,\text{mEq/L}$$

Infusing 0.9% NaCl (SID = 0) reduces the plasma SID → hyperchloraemic metabolic acidosis

Volumes > 30 mL/kg of normal saline can also cause hyperkalaemia by this mechanism and via tubular effects.

Balanced Solutions and Their Buffers

Lactated Ringer's, Hartmann's, and Ringer's Acetate contain organic anion buffers (lactate or acetate) that are metabolised to bicarbonate in the liver, thereby providing a physiological buffer effect without the acidosis of saline:

$$\text{Lactate}^- + \text{O}_2 \xrightarrow{\text{liver}} \text{HCO}_3^-$$ $$\text{Acetate}^- \xrightarrow{\text{liver/muscle}} \text{HCO}_3^-$$

However, large volumes of balanced solutions can produce:

• Hyperlactataemia (from lactate load in LR/Hartmann's), confounds lactate as a marker of hypoperfusion
• Metabolic alkalosis (excessive buffer generation)
• Hypotonicity (Na⁺ slightly lower than plasma in most balanced solutions)
• Cardiotoxicity from acetate load at high infusion rates

Plasmalyte A is the most physiologically balanced: Na⁺ 140, K⁺ 5, Cl⁻ 98, pH 7.4, with both acetate and gluconate as buffers, closely approximating plasma composition.

Calcium and Blood Product Interaction

Calcium-containing balanced solutions (e.g. Hartmann's, LR) may cause formation of microthrombi when co-infused with citrate-containing banked blood, clinically relevant during massive transfusion.

Glucose-Containing Crystalloids

Risks of glucose-containing solutions in ICU:

• Hyperglycaemia is common in critically ill patients due to the surgical stress response; iatrogenic glucose loading worsens this
• Hyperglycaemia induces osmotic diuresis, limiting resuscitation efficacy
• Aggravates ischaemic injury in brain and heart, avoid glucose-containing solutions in patients at risk of focal ischaemia
• D5W is effectively a hypotonic solution once glucose is metabolised, causes cerebral oedema in head-injured patients

Context-Sensitivity of Crystalloid Volume Effect

The volume effect of crystalloids is not fixed, it is context-sensitive:

• During haemorrhage or hypovolaemia: intravascular space accommodates fluid, and a greater fraction remains intravascular
• During normovolaemia: excess fluid is rapidly eliminated renally or redistributes to the interstitium
• In sepsis: glycocalyx disruption and increased capillary permeability cause crystalloids to distribute rapidly to the interstitium, worsening oedema without sustained plasma expansion

The glycocalyx, a proteoglycan layer lining the luminal endothelium, normally acts as a barrier to fluid movement. Rapid infusion can cause its deterioration. The revised Starling equation incorporating glycocalyx function:

$$J_v = K_f \left( P_v - P_i - \sigma\Pi_{esl} - \Pi_s \right)$$

Where $\Pi_{esl}$ is the oncotic pressure within the endothelial surface layer. In normal conditions, sub-glycocalyx oncotic pressure approaches zero, minimising net filtration.

Landmark trials, balanced vs saline

The mechanism-based rationale for balanced crystalloids (avoidance of hyperchloraemia, preservation of SID) is well established. Whether this translates to patient-centred outcomes has been tested in four major RCTs, collectively enrolling >45,000 patients.

SMART

Semler et al. NEJM 2018, Pragmatic cluster-crossover RCT; single-centre (Vanderbilt), 5 ICUs; n = 15,802 critically ill adults. Clinicians chose Plasma-Lyte A or Lactated Ringer's (balanced arm) vs 0.9% saline.

• Primary outcome: MAKE30 composite, death, new renal replacement therapy, or persistent renal dysfunction (creatinine ≥200% baseline) at 30 days.
• Result: Balanced 14.3% vs saline 15.4%; p = 0.04.
• Individual mortality components were not independently significant, the composite was driven by the renal signal, consistent with the Stewart SID mechanism: supraphysiological chloride load reduces SID → renal afferent vasoconstriction → AKI.
• Limitations: single-centre, unblinded clinician choice of balanced fluid, marginal p-value.

SALT-ED

Self et al. NEJM 2018, Same investigator group; ED non-ICU adults requiring IV fluid followed by ward admission; n = 13,347.

• Primary outcome: Hospital-free days to day 28, no difference.
• Secondary MAKE30: Balanced 4.7% vs saline 5.6%; p = 0.01.
• The AKI signal persisted but the clinical significance of hospital-free days was null, less critically ill population, smaller absolute fluid volumes administered.

PLUS

Finfer et al. NEJM 2022, Multicentre RCT; Australia and New Zealand (53 ICUs); n = 5,037 ICU adults expected to require IV fluid and an ICU stay ≥48 h. Balanced fluid was Plasma-Lyte 148 exclusively.

• Primary outcome: 90-day all-cause mortality.
• Result: 21.8% (balanced) vs 22.0% (saline); p = 0.90. Neutral.
• Persistent kidney dysfunction at 90 days: no difference.
• This is the largest ANZ-based ICU fluid RCT and directly contradicts the renal-mortality signal from SMART in a more rigorous design. Notably, PLUS ICUs had lower baseline chloride in the saline arm than historical estimates, clinicians may have been more conservative with saline volumes, diluting any treatment effect.

BaSICS

Zampieri et al. JAMA 2021, Multicentre 2×2 factorial RCT (balanced vs saline; slow vs fast infusion rate); Brazil (75 ICUs); n = 11,052 ICU adults.

• Primary outcome: 90-day all-cause mortality.
• Result: 26.4% (balanced) vs 27.2% (saline); p = 0.47. Neutral.
• Critical subgroup: Patients with traumatic brain injury had higher mortality with balanced solutions (p for interaction = 0.02).
• Mechanistic basis: Plasma-Lyte 148 contains Na⁺ 140 mmol/L vs saline Na⁺ 154 mmol/L, balanced solutions are slightly hypotonic relative to plasma and to saline; in a brain with impaired autoregulation, even modest free water gain promotes cerebral oedema and raised ICP.
• Practical implication: saline is preferred in TBI. This is now the prevailing ANZ/CICM position.

Key numbers

MAKE30 = death + new RRT + creatinine ≥200% baseline at 30 days. PL-A = Plasma-Lyte A; PL-148 = Plasma-Lyte 148; LR = Lactated Ringer's.

Bottom line

Pooling all four trials (~45,000 patients) the absolute mortality reduction with balanced crystalloids approximates 1%, not statistically significant once PLUS and BaSICS are included, and clinical equipoise exists for the average ICU patient. The renal signal from SMART and SALT-ED is biologically plausible (hyperchloraemia → reduced SID → renal vasoconstriction) but was not reproduced as a hard mortality endpoint in the larger ANZ and Brazilian trials. For most critically ill adults, the choice of balanced crystalloid vs saline is unlikely to determine survival; however, the quality of the safety signal around saline (hyperchloraemic acidosis, potential AKI) supports balanced solutions as the default choice in sepsis and general resuscitation. The single exception with unambiguous directionality is traumatic brain injury, where the relative hypotonicity of balanced solutions (Na⁺ 140 mmol/L vs 154 mmol/L) risks worsening cerebral oedema, saline remains preferred in this population. Calcium-containing balanced solutions (Hartmann's, LR) retain the co-infusion caution with citrated blood products regardless of trial outcomes.

Adverse Effects Summary

ICU Relevance

Fluid Selection in Critical Illness

Monitoring Targets When Administering Crystalloids

Standard ICU targets when using crystalloids for resuscitation include:

• MAP > 65-70 mmHg
• Urine output > 0.5 mL/kg/h
• CVP 10-15 mmHg (acknowledging its limitations as a static measure)
• Pulse pressure variation (PPV) or stroke volume variation (SVV), dynamic, superior indicators of fluid responsiveness compared with CVP or MAP alone

Avoiding Excess: The Cost of Liberal Crystalloid Therapy

Excessive crystalloid administration in critically ill patients has demonstrated harms:

• Pulmonary oedema: reduced total lung capacity, increased ventilator days
• Abdominal compartment syndrome: raised intraabdominal pressure with multi-organ failure
• Dilutional coagulopathy
• Cerebral oedema: particularly with hypotonic solutions or high-volume saline
• Confounding of lactate interpretation when using lactate-containing solutions

The surviving sepsis guidelines recommend 30 mL/kg of isotonic crystalloid for initial resuscitation in septic shock, but current evidence supports a move toward early vasopressor initiation (norepinephrine), smaller crystalloid volumes, and goal-directed, individualized fluid therapy using dynamic measures of fluid responsiveness.

Kinetic Principle for Clinical Practice

Because only ~20% of isotonic crystalloid remains intravascular, roughly 3-4 L of crystalloid must be infused to achieve a ~750 mL plasma volume expansion equivalent to 1 unit of packed red cells, at the cost of 2.25-3 L of interstitial oedema. This ratio reinforces the principle that crystalloids are appropriate for ECF replacement (e.g. diarrhoea, GI losses) but are inefficient for intravascular resuscitation, particularly in patients with disrupted capillary barriers such as in sepsis or burns.