Endothelial Glycocalyx and the Revised Starling Principle

Tissue fluid balance depends on how much fluid leaves the microcirculation and how effectively it returns. The revised Starling principle explains why the protein environment immediately beneath the endothelial surface matters. It also explains why raising plasma albumin does not automatically draw accumulated edema back into ordinary capillaries and keep it there.

16.1 The Glycocalyx Is Part of the Blood–Vessel Interface

The endothelial glycocalyx is a hydrated molecular layer on the blood-facing surface of endothelial cells. Proteoglycans have protein cores with attached carbohydrate chains. These molecules, together with hyaluronan and other components, form a matrix that interacts with water and plasma constituents.

The endothelial surface layer includes the glycocalyx and associated molecules from the circulation. It extends into the lumen, so circulating cells and proteins encounter it before reaching the endothelial cell membrane. It is a dynamic interface rather than an inert coating of uniform thickness.

Vink and Duling demonstrated that macromolecules, red cells, and leukocytes did not all have access to the same luminal region in mammalian capillaries. Their observations helped establish that the functional boundary of the blood compartment extends beyond the cell membrane. A vessel's anatomical lumen and the space accessible to a particular circulating component are therefore different concepts. 1

This surface participates in several functions. It influences the movement of water and macromolecules, limits direct contact between blood cells and the endothelial membrane, and helps transmit mechanical signals generated by flowing blood. Changes in one function can affect the others. For example, altered access to endothelial adhesion molecules can change leukocyte interaction with the vessel wall.

The glycocalyx works alongside endothelial junctions, the basement membrane, and tissue-specific structures. It is not the sole barrier to every substance. A change in a surface component therefore does not predict an identical change in water movement, protein passage, and cell adhesion.

Vascular beds also differ. Continuous systemic microvessels, renal glomerular vessels, pulmonary vessels, and hepatic sinusoids do not have interchangeable barriers. Their pressures and surrounding tissues differ as well. Findings from one preparation identify a mechanism, but do not supply a universal glycocalyx thickness or a permeability value for the entire patient.

For the critical care physician, the useful starting point is an active blood–vessel interface. Its condition helps determine how a given pressure and protein environment affects tissue fluid exchange. Understanding that interface makes the revised Starling principle easier to follow.

16.2 The Original Pressure–Oncotic Balance

Hydrostatic pressure favors fluid movement from higher toward lower pressure. Across a microvascular wall, capillary hydrostatic pressure generally favors movement toward the tissue, while tissue hydrostatic pressure opposes that movement when it is positive. Their difference provides the hydrostatic part of the exchange force.

Proteins provide an osmotic influence when the barrier restricts their passage. This is called colloid osmotic pressure, or oncotic pressure. If the effective protein concentration is greater on the blood side, the oncotic difference opposes outward fluid movement. The effect depends on how effectively the barrier restricts the proteins, not simply on their presence.

Starling's experiments established the importance of hydrostatic and protein osmotic influences in tissue fluid exchange. Those influences remain fundamental. The revision concerns where the effective protein difference acts and how the local protein environment changes as fluid moves. 2

The familiar classical equation uses capillary and interstitial hydrostatic pressures, plasma and bulk interstitial oncotic pressures, a filtration coefficient, and a protein reflection coefficient. The filtration coefficient combines hydraulic conductivity with exchange area. A larger value means more fluid crosses for the same net driving pressure, because the wall conducts water more readily, more surface is available, or both.

The reflection coefficient describes how effectively a particular solute generates an osmotic effect across that barrier. A value near one indicates strong restriction. A value near zero indicates little effective reflection and little sustained osmotic opposition from that solute. It is dimensionless and should not be read as a direct measurement of the percentage of glycocalyx remaining.

A common teaching diagram adds arterial-end filtration followed by venous-end reabsorption. Capillary pressure falls along the vessel, so this appears to follow naturally if the protein pressures are held fixed. The difficulty is that the tissue-side protein environment does not remain fixed when filtration decreases or reverses.

The important distinction is between an immediate response and a steady state. A pressure change can initially produce absorption, yet the local conditions can then change until that absorption diminishes. The revised principle retains the pressure balance while explaining why the later response may differ from the first one.

16.3 The Subglycocalyx Protein Environment

Bulk interstitial fluid is the fluid surrounding the tissue cells. It need not have the same protein concentration as the fluid immediately beneath the selective endothelial surface layer. This local region on the tissue side of the glycocalyx is called the subglycocalyx space.

During filtration, fluid crosses the selective layer while much of the plasma protein is restricted. The filtrate is consequently relatively protein-poor. As it passes through the local pathway, it helps maintain a low protein concentration immediately beneath the barrier.

Interstitial protein can diffuse toward that region from the tissue side. Outward fluid movement and the geometry of the endothelial pathway limit how readily it accumulates there. The local concentration therefore reflects competing processes: delivery by diffusion, removal with outward fluid movement, and the barrier's restriction of protein passage.

The effective oncotic difference is between plasma and this local subglycocalyx fluid. Subtracting bulk interstitial oncotic pressure from plasma oncotic pressure can give a different answer because the bulk sample does not describe the protected local environment beneath the barrier.

Adamson and colleagues tested this concept in nonfenestrated rat microvessels. They combined measurements of oncotic forces with analysis of the endothelial pathway. Their findings supported a lower protein concentration beneath the glycocalyx than in the surrounding interstitium during filtration. This provides experimental support for the local-gradient model, not a method for measuring the same compartment at the bedside. 3

Interstitial protein still matters. It supplies protein that can reach the local space and contributes to the tissue and lymphatic environment. The revision does not remove the interstitium from the system. It distinguishes the bulk compartment from the immediate tissue side of the selective barrier.

Nor is subglycocalyx protein permanently low. When outward filtration falls, less protein-poor fluid passes through the pathway. Protein arriving from the tissue can then raise the local concentration. The resulting change in the oncotic difference is central to understanding why sustained absorption is often limited.

16.4 Writing the Revised Relationship

A simplified form of the revised relationship is:

Jv is net fluid movement. A positive outward value represents filtration toward tissue. Pc is capillary hydrostatic pressure, and Pi is the relevant tissue-side hydrostatic pressure. Their difference is the hydrostatic driving term.

The symbols πp and πg represent plasma and subglycocalyx oncotic pressures. Subtracting πg from πp gives the local oncotic difference. Multiplying by the protein reflection coefficient, σ, adjusts for the barrier's effective restriction of that protein. Kf is the filtration coefficient, which converts the net pressure term into a fluid movement rate.

All pressure terms must use the same units. If pressures are expressed in mm Hg and Kf in mL/min/mm Hg for the vascular bed being modeled, Jv is expressed in mL/min. The reflection coefficient has no units. Without an appropriate Kf, the bracket gives a pressure difference, not a measured volume of fluid leaving the circulation.

Consider illustrative values of Pc = 20 mm Hg and Pi = 0 mm Hg. These represent pressure inside the capillary and on the tissue side. Their difference is:

20 − 0 = 20 mm Hg outward hydrostatic term

Now suppose plasma oncotic pressure is 25 mm Hg and subglycocalyx oncotic pressure is 5 mm Hg. Their difference is 20 mm Hg. With a reflection coefficient of 0.9, the effective oncotic opposition is:

0.9 × (25 − 5) = 18 mm Hg

Subtracting that opposition from the hydrostatic term gives:

20 − 18 = 2 mm Hg net outward term

If capillary pressure rises to 30 mm Hg while the other assumed values remain fixed, the hydrostatic term becomes 30 mm Hg. The effective oncotic opposition remains 18 mm Hg, so the net outward term becomes:

30 − 18 = 12 mm Hg

The example shows how a 10 mm Hg rise in capillary pressure increases the calculated net outward term from 2 to 12 mm Hg. That is a sixfold increase in this fixed-parameter model. It does not predict a sixfold steady-state rise in a patient's filtration. Local protein concentration, tissue pressure, exchange area, and hydraulic conductivity can change during the response.

A lower reflection coefficient reduces oncotic opposition for the same protein difference. A higher Kf increases fluid movement for the same net pressure term. Injury can change both, so edema cannot identify which component is responsible. Routine arterial or central venous pressure also cannot be substituted directly for capillary pressure in every organ.

16.5 Why Venous-End Reabsorption Is Often Transient

Suppose capillary pressure suddenly falls while the protein environment initially remains unchanged. The outward hydrostatic term decreases immediately. If oncotic opposition now exceeds that term, the initial net movement can reverse and fluid enters the vessel from the tissue.

The conditions supporting absorption then change. Reduced outward filtration means less protein-poor fluid passes beneath the glycocalyx. Protein arriving from the tissue side becomes more influential, and the local protein concentration rises. This reduces the plasma-to-subglycocalyx oncotic difference that initially favored absorption.

The balance therefore moves away from the initial absorbing state. In many continuous microvascular beds, the resulting steady state has little net movement or low filtration, rather than persistent absorption. A prediction based on the initial protein difference will overestimate how long absorption continues if that difference is held artificially constant.

Michel and Phillips compared immediate and steady-state responses to pressure changes in perfused frog mesenteric capillaries. At lower pressures, the immediate measurements showed absorption, whereas steady-state measurements showed very small movement or slight filtration. Their experiment demonstrates why the first response cannot simply be extended indefinitely. 4

Transient absorption remains physiologically relevant. An abrupt pressure reduction during volume loss can initially move fluid into the circulation. The revised principle does not prohibit this response. It explains why its magnitude and duration depend on how the local exchange environment adjusts.

The same distinction applies to increasing plasma oncotic pressure. A larger initial gradient may reduce filtration or favor transient absorption, but it does not guarantee sustained removal of established edema through ordinary venous capillaries. Hydrostatic pressure, protein movement, barrier function, and lymphatic return continue to determine the later state.

This is why edema clearance requires more than a plasma albumin measurement. The return pathway and the pressures maintaining filtration must also be considered. Lymphatic transport is central when ordinary capillaries do not provide sustained net absorption.

16.6 Scope, Exceptions, and Specialized Capillary Beds

The revised model is particularly useful for continuous, relatively protein-restricting microvessels. It should not be converted into an absolute claim that no capillary can absorb fluid. Different organs operate with different barriers, pressures, and sources of tissue fluid.

Renal peritubular capillaries illustrate this distinction. Tubular reabsorption continuously transfers fluid from the nephron into the surrounding interstitium. The peritubular circulation helps return that fluid to blood. Its exchange environment therefore includes an ongoing source of fluid that is absent from a simple resting-muscle model.

Brenner and colleagues examined peritubular protein concentration and proximal tubular reabsorption in rats. Changes in peritubular oncotic conditions were associated with changes in proximal reabsorption, while later responses also required other mechanisms. The experiment demonstrates the importance of the local renal environment rather than a universal rule that raising plasma protein increases reabsorption everywhere. 5

Fenestrated and sinusoidal vessels differ from continuous systemic vessels in their pathways for water and protein. Endothelial junctions, basement membranes, and specialized tissue barriers also contribute. A filtration coefficient or reflection coefficient measured in one preparation cannot be assigned unchanged to another organ.

Time matters as well as anatomy. Abrupt blood loss, stable tissue fluid balance, established inflammatory edema, and chronic venous hypertension describe different conditions. An immediate pressure response in an isolated vessel does not establish the sustained behavior of an injured organ.

When applying the model, identify the vascular bed and ask whether the state is changing or approximately stable. Then identify the relevant hydrostatic pressures, protein compartments, and fluid return pathways. This preserves the model's explanatory value without making its scope broader than the evidence supports.

16.7 Lymphatic Return Completes the Fluid Balance

Filtered fluid and protein must leave the interstitium if tissue volume is to remain stable. Lymphatic vessels provide a major return route to the venous circulation. Their importance follows directly from the limited sustained absorption in many ordinary capillary beds.

Fluid enters initial lymphatic vessels from the tissue. It then passes through collecting vessels whose valves and contractile segments help move it forward. Tissue movement and surrounding pressure changes can assist transport. The lymph ultimately returns both fluid and protein to blood.

When filtration rises, lymphatic transport can increase and buffer tissue accumulation. Increasing interstitial pressure can also oppose further filtration under some conditions. These responses provide reserve, but neither can compensate indefinitely for excessive entry or impaired drainage.

Edema develops when fluid enters faster than it can be removed. Tissue structure determines how the accumulating volume affects pressure and function. A compliant region can accommodate volume differently from a constrained organ. The same total fluid balance therefore does not produce the same regional consequences in every patient.

Venous congestion can impair both sides of the balance. Higher downstream vascular pressure can raise microvascular hydrostatic pressure and favor filtration. Higher pressure where lymph enters the venous system can also make drainage less favorable. The magnitude depends on the regional circulation, surrounding pressure, and lymphatic pumping function.

Barrier injury adds fluid and protein to the interstitium, increasing what drainage must remove. If lymphatic transport is already impaired, even a modest additional filtration load can produce accumulation. Serum albumin alone cannot describe this balance because it measures neither local filtration nor lymph flow.

Edema clearance consequently depends on reducing excessive entry and maintaining effective return. Depending on the clinical cause, this may require relief of congestion, restoration of renal fluid excretion, and time for drainage and tissue recovery. The later chapters examine interstitial pressure, lymphatic transport, and plasma refill in more detail.

16.8 Glycocalyx Injury: Mechanisms and Human Observations

Inflammation and ischemia–reperfusion can alter the endothelial surface. Enzymes can cleave surface molecules and release their fragments. Oxidative and inflammatory processes can also affect endothelial cells and junctions. The resulting disturbance involves interacting structures rather than removal of a uniform protective sheet.

Schmidt and colleagues studied the pulmonary glycocalyx during experimental sepsis. Heparanase-mediated cleavage of heparan sulfate exposed endothelial adhesion sites and promoted neutrophil interaction. Their experiments linked these changes to inflammatory lung injury and included supporting observations in human disease. They identify a mechanism, but do not establish a glycocalyx-directed treatment that improves outcomes in patients with sepsis. 6

Rehm and colleagues measured circulating glycocalyx-related components during major vascular surgery with ischemia–reperfusion exposures. Their human findings supported shedding during vascular stress. Such measurements demonstrate a circulating response; they do not measure the permeability of every organ or establish how much of its surface layer remains functional. 7

Injury can influence fluid exchange through more than one route. Increased hydraulic conductivity permits more water movement for a given net pressure. Reduced protein restriction weakens effective oncotic opposition. Increased leukocyte adhesion and inflammatory signaling can further disturb endothelial function. Which effect predominates depends on the vessel and disease process.

Edema alone cannot identify glycocalyx injury. Elevated microvascular pressure, impaired drainage, renal sodium retention, and low plasma protein can also contribute. In critical illness, these mechanisms commonly overlap, so assigning all swelling to a single surface lesion obscures the rest of the assessment.

Treatment should address demonstrated causes such as infection, ischemia, obstruction, and congestion. Molecular mechanisms can improve the explanation of the disease, but an appealing explanation is not itself evidence for a specific drug, fluid, or biomarker target. Clinical benefit requires separate evaluation.

16.9 What Shedding Markers and Volume-Loading Studies Can Establish

Syndecan-1, heparan sulfate, and hyaluronan are studied as circulating signals of surface-layer disturbance. Their concentration depends on how much enters the circulation, the volume in which it distributes, and how rapidly it is cleared. A concentration is not a direct measurement of endothelial thickness or hydraulic function.

The source also matters. These molecules are not exclusive to one endothelial surface, and release from other tissues can contribute. Kidney and liver function can influence their handling. A higher concentration cannot therefore be converted directly into a precise amount of whole-body glycocalyx loss.

Chappell and colleagues compared volume loading with acute normovolemic hemodilution in a small perioperative study. The patients were selected surgical patients, and the protocol used a particular hydroxyethyl starch preparation. Volume loading was accompanied by increased atrial natriuretic peptide and some glycocalyx-related markers. The study did not test routine crystalloid boluses in septic shock. 8

Replacing an actual circulatory deficit and producing unnecessary hypervolemia are different interventions. The study supports attention to that distinction. It does not establish that every clinically necessary fluid bolus causes important glycocalyx damage, or that withholding needed volume improves outcomes.

Atrial stretch can stimulate natriuretic peptide release, providing a plausible link between excess filling and endothelial changes. However, observing peptide and shedding-marker increases together does not isolate every step of that causal pathway in an individual patient. The clinical effects still depend on perfusion, pressure, congestion, and the underlying illness.

The same standard applies to a proposed protective treatment. A lower marker concentration is a surrogate result. It could reflect less release, altered distribution, or changed clearance. Meaningful barrier, organ, or clinical outcomes are needed before treating it as proof of benefit.

At present, these observations do not supply a validated biomarker-based fluid dose or stopping point. Bedside decisions still require a supported reason for more volume, an assessment of fluid tolerance, and reassessment of the actual response.

16.10 Implications for Crystalloid, Albumin, and Edema

An infused fluid enters a changing circulation. It can alter filling, capillary pressure, plasma dilution, and urine production. At the same time, barrier function and lymphatic return influence its distribution. The proportion remaining intravascularly therefore depends on the patient and the time of measurement.

Crystalloid can increase stroke volume when additional filling increases forward flow. Its water and electrolytes then redistribute and undergo renal handling. A single fixed intravascular fraction cannot describe every infusion in every shock state or at every later time point.

Albumin can expand plasma volume and increase plasma oncotic pressure. The revised principle does not remove these effects. It explains why their clinical consequences depend on the effective barrier, local protein conditions, and hydrostatic response. Plasma expansion and sustained clearance of tissue edema are different outcomes.

ALBIOS compared albumin replacement plus crystalloid with crystalloid alone in severe sepsis. The albumin group had differences in arterial pressure and net fluid balance, but no significant overall mortality difference at 28 or 90 days. The trial tested a clinical replacement strategy, not the revised Starling model in isolation. 9

The 2026 ARISS trial tested albumin replacement in septic shock. It randomized 440 patients and did not establish a significant 90-day mortality benefit. Recruitment stopped early because enrollment was slow, leaving important uncertainty. This result does not establish equivalence or justify universal conclusions about every indication for albumin. 10

Table 16.1. Practical interpretations of the revised principle.

Fluid choice requires the diagnosis, current hemodynamics, applicable treatment evidence, and tolerance of added volume. The exchange model helps predict mechanisms to watch. It cannot choose a resuscitation fluid independently or turn a low albumin concentration into proof that replacement will correct shock.

16.11 A Bedside Interpretation of Hypoalbuminemia and Edema

Consider an illustrative patient with septic shock, hypoalbuminemia, generalized edema, and hypotension after initial resuscitation. Venous pressure has risen. A technically adequate reversible filling test produces no clear stroke-volume increase.

First explain the edema without assigning it a single cause. Inflammation can alter protein restriction and hydraulic conductivity. Elevated venous pressure can favor filtration and impede drainage. Renal sodium and water retention can maintain extracellular expansion. Low plasma albumin modifies the oncotic environment within this larger system.

The low albumin concentration does not show that replacement will return the accumulated edema to blood or correct the hypotension. Even if plasma volume expands, forward flow may not improve. Capillary pressure can rise, and the local protein environment can adjust. These are separate responses that require observation.

The filling test weakens the expectation that more volume will increase stroke volume at that time. Rising venous pressure increases concern about tolerance. Together, these findings make a further infusion difficult to justify solely as a means of improving output. They do not establish the cause of the continuing hypotension.

Reassess vascular tone, ventricular performance, rhythm, and possible obstruction. Confirm that the measurements and filling test were appropriate for the clinical conditions. Review infection treatment and source control. The pressure may require treatment, but the absence of a useful flow response argues against assuming that inadequate volume is the explanation.

If a fluid is considered for a separate supported reason, state the intended benefit and assess whether it occurs. Follow stroke volume, perfusion, pressure, congestion, fluid balance, and organ function. A rise in plasma albumin or filling pressure alone is insufficient evidence of a useful resuscitation response.

The revised principle replaces a static picture of fluid leaving and returning along one capillary with a dynamic exchange system. Hydrostatic pressures, local protein, barrier properties, and drainage interact over time. Its bedside value is to make those mechanisms explicit while keeping the cause of shock and the observed treatment response central.

References

1. Vink H, Duling BR. Identification of distinct luminal domains for macromolecules, erythrocytes, and leukocytes within mammalian capillaries. Circ Res. 1996;79(3):581–589. doi:10.1161/01.RES.79.3.581.

2. Starling EH. On the absorption of fluids from the connective tissue spaces. J Physiol. 1896;19(4):312–326. doi:10.1113/jphysiol.1896.sp000596.

3. Adamson RH, Lenz JF, Zhang X, Adamson GN, Weinbaum S, Curry FE. Oncotic pressures opposing filtration across non-fenestrated rat microvessels. J Physiol. 2004;557(Pt 3):889–907. doi:10.1113/jphysiol.2003.058255.

4. Michel CC, Phillips ME. Steady-state fluid filtration at different capillary pressures in perfused frog mesenteric capillaries. J Physiol. 1987;388:421–435. doi:10.1113/jphysiol.1987.sp016622.

5. Brenner BM, Falchuk KH, Keimowitz RI, Berliner RW. The relationship between peritubular capillary protein concentration and fluid reabsorption by the renal proximal tubule. J Clin Invest. 1969;48(8):1519–1531. doi:10.1172/JCI106118.

6. Schmidt EP, Yang Y, Janssen WJ, et al. The pulmonary endothelial glycocalyx regulates neutrophil adhesion and lung injury during experimental sepsis. Nat Med. 2012;18(8):1217–1223. doi:10.1038/nm.2843.

7. Rehm M, Bruegger D, Christ F, et al. Shedding of the endothelial glycocalyx in patients undergoing major vascular surgery with global and regional ischemia. Circulation. 2007;116(17):1896–1906. doi:10.1161/CIRCULATIONAHA.106.684852.

8. Chappell D, Bruegger D, Potzel J, et al. Hypervolemia increases release of atrial natriuretic peptide and shedding of the endothelial glycocalyx. Crit Care. 2014;18:538. doi:10.1186/s13054-014-0538-5.

9. Caironi P, Tognoni G, Masson S, et al.; ALBIOS Study Investigators. Albumin replacement in patients with severe sepsis or septic shock. N Engl J Med. 2014;370(15):1412–1421. doi:10.1056/NEJMoa1305727.

10. Sakr Y, Nierhaus A, Schumacher U, et al.; SepNet Critical Care Trials Group and ARISS investigators. Albumin replacement therapy in septic shock: a randomized clinical trial. JAMA Netw Open. 2026;9(2):e2559297. doi:10.1001/jamanetworkopen.2025.59297.