The Microcirculation: Capillary Perfusion and Oxygen Exchange

An adequate arterial pressure and cardiac output do not ensure that oxygen reaches every tissue region. Blood must enter the capillary network, reach the cells that need it, and release oxygen across the exchange pathway. Failure at any of these steps can leave tissue hypoxic despite reassuring measurements in the larger circulation.

15.1 The Microcirculation Is a Network With Several Tasks

The microcirculation comprises small resistance vessels, capillaries, and venules. These vessels work together to distribute blood through a tissue and return it to the larger veins. Their arrangement also determines where exchange occurs and how closely flowing blood approaches the cells it supplies.

Arterioles regulate much of the resistance to local inflow. Contraction of their smooth muscle narrows the lumen and increases resistance. Relaxation usually lowers resistance, allowing more flow when the driving pressure is maintained. Arteriolar resistance also influences how much arterial pressure reaches the downstream exchange vessels.

Capillaries provide an extensive surface across which oxygen, nutrients, and other substances can move. Their walls differ between organs. A continuous endothelial barrier restricts passage differently from a fenestrated surface. Capillary exchange therefore depends on the properties of the vessel wall as well as the blood flowing through it.

Venules collect blood from the capillary network. They are also important sites of leukocyte interaction with the vessel wall during inflammation. Adhesion and changes in endothelial permeability can alter the local circulation even when the larger vessels remain open.

Local control is not identical in every tissue. Pericytes and other supporting cells influence microvascular behavior in selected organs. Their contribution depends on the vessel segment and tissue. A model in which every capillary has the same independent muscular valve does not represent this diversity.

The network contains many parallel pathways. When one pathway becomes obstructed or poorly perfused, additional blood entering the organ may travel through other, more accessible pathways. Total organ inflow can then rise without restoring red-cell passage beside the affected cells.

For this reason, assess the vessels together with the tissue they supply. An open vessel is useful for oxygen delivery only if oxygenated red cells pass through it and oxygen can reach nearby cells. Network structure, blood-cell movement, and tissue metabolism determine whether apparently adequate inflow produces effective exchange.

15.2 Red Cells Change the Behavior of Small-Vessel Flow

Blood contains cells suspended in plasma. In a large vessel, a sample can describe their average concentration reasonably well. In a narrow vessel, individual cells occupy a substantial part of the lumen. Their size, deformability, and position then have a greater influence on flow.

Red cells deform as they pass through narrow channels. Their distribution across the vessel is also uneven, with a relatively cell-poor layer near the wall. At a branch, red cells and plasma may divide unequally between the daughter vessels. Two neighboring pathways can therefore have different local hematocrits even though both receive blood from the same upstream vessel.

Fåhraeus and Lindqvist studied blood flowing through narrow glass tubes. They showed that its apparent viscosity depends on tube diameter. This established that blood cannot always be treated as a uniform liquid with one fixed viscosity. The experiments used a simplified preparation; living vessels add branching, deformable walls, and interactions with the endothelium. 1

Apparent viscosity describes the viscosity inferred from the observed pressure–flow relationship. It reflects how the suspension behaves under those conditions. A change in apparent viscosity does not mean that narrowing a vessel necessarily lowers its resistance. The geometric effect of a smaller lumen and the changing behavior of the cells both influence the final resistance.

Inflammation can disturb several parts of this system. Less deformable red cells pass less easily through narrow pathways. Adherent leukocytes and platelet interactions can impede passage. Microthrombi can obstruct individual channels. The result may be intermittent or stopped red-cell flow beside vessels that continue to carry substantial blood.

This distinction matters because hemoglobin carries most blood oxygen. Plasma movement through a pathway with few red cells delivers less oxygen than the same volume of blood with a higher red-cell concentration, if hemoglobin saturation is otherwise comparable. Seeing fluid move through a vessel does not establish how much oxygen it carries.

Local oxygen delivery thus depends on red-cell flow and hemoglobin oxygen content, not simply on the volume of fluid passing through the network. This helps explain why a large-vessel hematocrit or cardiac output cannot fully describe oxygen supply at each exchange site.

15.3 Convection Delivers Oxygen; Diffusion Transfers It to Cells

Convection transports oxygen with moving blood. It brings oxygenated red cells into a tissue region. Diffusion then moves oxygen from blood toward tissue down an oxygen partial-pressure gradient. Effective supply requires blood to arrive and oxygen to cross the distance between that blood and the cells.

Most oxygen in arterial blood is bound to hemoglobin. Before it can diffuse through plasma and tissue, it must leave hemoglobin and enter the dissolved phase. As dissolved oxygen leaves the blood, hemoglobin releases more. Hemoglobin therefore supplies a large oxygen store while the dissolved oxygen establishes the partial pressure that drives diffusion.

Oxygen content and oxygen tension describe different properties. Content is the amount of oxygen carried in a volume of blood. Tension is its partial pressure. A patient with anemia can have a satisfactory arterial oxygen tension but carry less oxygen in each volume of blood because less hemoglobin is available.

The main influences on diffusion can be shown with a simplified relationship:

Exchange area is the surface available for oxygen transfer. The pressure difference is between the blood and the tissue across the exchange pathway. Diffusion distance is the effective length of that pathway. Greater area or a larger pressure difference favors transfer. A longer distance reduces transfer if the other conditions remain unchanged.

For example, suppose diffusion distance doubles. It changes from one unit of distance to two units. Because distance is in the denominator, the calculated transfer becomes one divided by two, or half its original value. This example assumes unchanged exchange area, oxygen pressure difference, and diffusion properties. It describes the relationship, not an exact prediction of what edema will do in a patient.

Krogh examined capillary distribution in muscle and calculated the oxygen pressure needed to supply surrounding tissue. His work established why the distance from a perfused capillary matters. The familiar cylindrical model is useful for teaching this principle, although actual tissues have more complex networks, variable oxygen consumption, and overlapping sources of supply. 2

Increasing global oxygen delivery primarily addresses convection. It may help when too little oxygenated blood enters the tissue. It cannot automatically correct a long diffusion distance or an area that blood does not reach. Conversely, restoring flow after prolonged ischemia does not ensure immediate cellular recovery, because injured cells may still be unable to use oxygen normally.

15.4 Functional Capillary Density and the Exchange Surface

Anatomical capillary density describes the vessels that exist in a tissue. Functional capillary density concerns the capillaries carrying red cells under the conditions of observation. A vessel can remain anatomically present while contributing little to oxygen delivery because red-cell passage has stopped.

Loss of perfused pathways changes the distribution of supply. The tissue beside a stopped capillary must obtain oxygen from another flowing vessel, if one is close enough. The effective distance from blood to those cells may increase. At the same time, less perfused exchange surface remains available within that tissue region.

Sending more blood through the remaining channels does not necessarily reverse these changes. Red-cell velocity can rise in those channels while the poorly supplied region remains distant from flowing blood. The average flow may look better even though the distribution of exchange pathways has not recovered.

The term recruitment often describes increased participation of the microvascular network. A previously weakly perfused pathway may begin carrying red cells more consistently. This does not prove that its lumen was completely closed beforehand. Changes in pressure, resistance, and the division of red cells between branches can alter its functional contribution without creating a new vessel.

Density and flow quality must therefore be considered separately. A dense network with sustained red-cell passage differs from the same visible network with intermittent flow. The proportion of vessels that are perfused also differs from their absolute density. A high proportion can still describe a sparse network if few vessels are visible in the sampled area.

Tissue demand determines whether the available network is sufficient. A pattern that supports resting metabolism may become inadequate when oxygen consumption rises or blood oxygen content falls. Functional density is an observation made in a particular physiological state, rather than a permanent characteristic of the patient.

These distinctions help interpret microvascular images. Density describes the available network; continuity and velocity describe aspects of passage through it. Their combined meaning depends on whether oxygenated red cells reach the tissue regions that need them.

15.5 Heterogeneous Flow Can Impair Oxygen Extraction

Heterogeneous flow means that neighboring pathways carry different amounts of blood. Some variation is normal. The concern is a distribution in which substantial flow passes through selected channels while other regions receive little or none. The average flow then conceals important local differences.

Consider a region beside a stopped capillary. Increasing flow in a more distant channel may deliver more oxygen to the organ as a whole, but that oxygen must still diffuse far enough to reach the undersupplied cells. If distance and demand prevent this, those cells remain vulnerable. Oxygen cannot be assumed to redistribute freely throughout the entire organ.

Transit time is the time blood spends in an exchange pathway. It depends on flow relative to the volume of blood within that pathway. A shorter transit time is not inherently harmful. Healthy tissues can support greater flow through coordinated changes in distribution, oxygen gradients, and exchange. Rapid passage becomes a problem when oxygen transfer cannot keep pace with local need or when it accompanies loss of other useful pathways.

An experimental sepsis study by Ellis and colleagues illustrates an important distinction. In rat skeletal muscle, more capillaries developed stopped flow. The capillaries that maintained normal flow extracted more oxygen. This was a compensatory response in the remaining flowing pathways, rather than evidence that extraction fell in every capillary. The experiment demonstrated how uneven red-cell delivery can impose greater demands on the pathways that remain perfused. 3

Individual capillaries and the whole tissue therefore need not show the same pattern. Increased extraction in one flowing pathway does not establish adequate oxygenation beside another pathway that has stopped. Conversely, observing maldistribution does not prove that mitochondria everywhere have lost the ability to use oxygen.

Venous blood mixes the blood leaving different tissue regions. Well-supplied pathways may return blood with substantial oxygen remaining, while poorly supplied tissue contributes little flow to the sample. A normal or high venous oxygen saturation can therefore coexist with regional hypoxia.

The same saturation can also reflect high delivery relative to demand or reduced cellular oxygen use. Its value alone cannot distinguish these explanations. Interpret it alongside the clinical circulation, hemoglobin, oxygenation, metabolic demand, and organ findings.

15.6 Edema and Barrier Injury Change the Exchange Environment

The exchange pathway includes the endothelial surface and the surrounding interstitium. Endothelial barrier injury allows more fluid and protein to leave the vascular space. Higher microvascular hydrostatic pressure can increase filtration, while inadequate lymphatic drainage limits removal. Tissue fluid accumulates when entry exceeds drainage.

Edema can separate cells from effectively perfused blood and increase the distance oxygen must cross. Its distribution matters as much as its total amount. Fluid accumulating near vulnerable cells can alter their supply even when whole-body fluid balance gives little information about that particular region.

In tissues with limited room to expand, accumulating fluid can also raise interstitial pressure. This may compress susceptible small vessels and impede flow. Compression then reduces convective supply as well as changing the diffusion pathway. The importance of each effect depends on organ structure and the pressures inside and outside the vessels.

The endothelial surface layer helps regulate permeability and interactions between blood cells and the vessel wall. Injury can disturb both fluid exchange and vascular signaling. The next chapter develops these mechanisms through the glycocalyx and the revised Starling principle.

Additional fluid can consequently have competing effects. If it raises stroke volume, it may increase convective oxygen delivery. If it also raises venous and microvascular pressures or escapes into the interstitium, it can worsen congestion and edema. Demonstrating fluid responsiveness does not establish that the tissue will tolerate the added volume.

Cellular oxygen use is a further, separate step. Mitochondrial dysfunction can limit the use of oxygen that has already reached the cell. This differs from inadequate inflow or impaired diffusion. The mechanisms can coexist, but an abnormal lactate or persistent organ dysfunction cannot identify which one predominates.

The bedside assessment therefore examines the sequence. Is enough oxygenated blood entering? Is it distributed near the affected cells? Can oxygen cross the exchange pathway? Is established cellular injury limiting recovery? No single routine value answers all four, so treatment depends on the supported explanation and its response over time.

15.7 Microcirculatory Dysfunction in Sepsis

Sepsis can disrupt endothelial signaling, blood-cell passage, coagulation, and local vascular responses. These changes need not affect every channel equally. Some vessels continue to carry blood while others develop slow, intermittent, or stopped flow. Restoring arterial pressure does not necessarily remove these local disturbances.

De Backer and colleagues demonstrated abnormal sublingual microvascular flow in patients with sepsis. Their observations showed why systemic measurements alone cannot describe the local exchange circulation. They measured the accessible vessels beneath the tongue, rather than directly measuring the microcirculation of every internal organ. 4

Sakr and colleagues followed patients with septic shock. Persistent microcirculatory abnormalities were associated with organ failure and death. This establishes prognostic association, but the observational design cannot show that treatment directed at an image-derived index improves survival. A marker can track disease severity without being a proven treatment target. 5

The term loss of hemodynamic coherence describes a mismatch between systemic and microvascular recovery. Arterial pressure or cardiac output improves, but local perfusion does not improve in parallel. The term identifies the mismatch; it does not explain whether its cause is obstruction, altered regulation, edema, or another mechanism.

Systemic resuscitation remains necessary when driving pressure or cardiac output is inadequate. A capillary network cannot compensate indefinitely for insufficient upstream supply. The additional task is to examine whether correcting the systemic limitation also improves tissue perfusion and organ function.

Persistent abnormalities should prompt that reassessment. Prolonged capillary refill, elevated lactate, and oliguria each provide useful information, but none directly establishes capillary obstruction. Temperature and vascular tone affect refill. Lactate depends on production and clearance. Urine output depends on filtration, tubular function, hormonal responses, and obstruction as well as perfusion.

The clinician needs an explanation that fits the measurements and their time course. Persistent organ dysfunction after restored pressure can reflect unresolved perfusion problems, established tissue injury, or both. Repeatedly increasing the same treatment without evidence of benefit does not resolve that distinction.

15.8 What Sublingual Microvascular Imaging Measures

Handheld sublingual microscopy uses the optical properties of hemoglobin to visualize red cells in small vessels. The image shows local vessel patterns and red-cell movement. It adds information that an arterial waveform or cardiac output measurement cannot provide, but its conclusions depend on the quality and location of the recording.

Different indices describe different features. Vessel density concerns how much vessel length is visible within the sampled area. Perfused vessel density concerns the portion carrying blood. The proportion of perfused vessels describes how much of the visible network is perfused. Flow indices and comparisons between fields describe continuity and heterogeneity.

Some measures are semiquantitative scores rather than direct measurements of volumetric flow. None independently measures cellular oxygen consumption. An image of red-cell passage does not show the oxygen gradient from each vessel to every surrounding cell or establish whether those cells can use the oxygen delivered.

The 2018 ESICM task-force consensus addressed acquisition and interpretation of sublingual images. It emphasized suitable focus, illumination, stability, and sampling, together with avoidance of excessive probe pressure. This is a methodological consensus, not an outcome trial proving that routine image-guided treatment benefits patients. 6

Probe pressure is especially important. Pressing on the tissue can reduce flow in the vessels being examined. The instrument can then create the abnormality attributed to the patient. Secretions, movement, poor focus, and selection of an unrepresentative field can also distort the assessment.

Sampling several suitable areas reduces dependence on one local view. Serial comparisons require similar acquisition conditions; otherwise, a change in technique may be mistaken for a physiological response. A technically inadequate clip should be rejected rather than converted into a confident diagnosis.

Finally, the sublingual site is one vascular bed. Its response may differ from the kidney, bowel, or brain because local anatomy and regulation differ. Improved sublingual flow supports a local observation. Broader conclusions require the systemic circulation, other perfusion signs, and organ findings to support them.

15.9 Peripheral Perfusion Assessment and Clinical Trials

Capillary refill time is the interval between releasing standardized pressure on a skin site and the return of its color. Compression displaces blood from the examined area. The return of color gives an accessible indication of peripheral perfusion under the measurement conditions.

Technique influences the observation. The site, applied pressure, duration of compression, lighting, and timing method should be consistent. Temperature, skin characteristics, and vascular tone also affect the result. A cold hand during systemic hypothermia cannot be interpreted in the same way as a similarly cold hand during otherwise warm septic shock.

Mottling and skin temperature gradients add related information. They describe the peripheral circulation rather than directly measuring blood flow through internal organs. Repeating the examination after an intervention can help establish whether the peripheral response changes in the expected direction.

In the 2019 ANDROMEDA-SHOCK trial, a resuscitation strategy targeting capillary refill was compared with a strategy targeting lactate in septic shock. The primary comparison of 28-day mortality did not reach conventional statistical significance. The trial did not establish a survival benefit or prove equivalence between the strategies. 7

ANDROMEDA-SHOCK-2 subsequently tested a personalized hemodynamic protocol targeting capillary refill against usual care. Its primary outcome ranked patients using death, duration of vital support, and hospital stay. The 2025 trial favored the protocol on that hierarchical outcome, mainly through shorter vital support. Mortality alone did not differ significantly. 8

The intervention included assessment of fluid responsiveness and other hemodynamic findings, with treatments selected and reassessed through an algorithm. The result therefore applies to the complete strategy. It cannot establish that measuring refill time alone produces the same benefit or that every prolonged value requires additional fluid.

At the bedside, peripheral perfusion signs provide another observation to integrate with systemic hemodynamics and organ function. Their usefulness increases when technique is consistent and changes are interpreted in context. An abnormal sign identifies a reason to reassess; the rest of the examination determines which treatment, if any, addresses its cause.

15.10 Why More Flow or More Vasodilation May Not Solve the Problem

Increasing cardiac output helps when inadequate inflow limits tissue supply. It may provide little benefit to a region that remains bypassed or obstructed. It also cannot directly reverse a diffusion limitation or restore injured mitochondria. The systemic response must therefore be interpreted through its tissue consequences.

Ospina-Tascón and colleagues examined microvascular changes after fluid administration in severe sepsis. In this nonrandomized study, perfusion improved in patients assessed early more than in those assessed later. The microvascular response was not explained simply by changes in global hemodynamics. These observations support reassessment of the response; they do not establish a universal timing rule or justify repeated fluid without regard to tolerance. 9

Vasodilation has a plausible local mechanism: lowering arteriolar resistance can increase flow if driving pressure is maintained. Systemic vasodilation can also lower that pressure. If the pressure fall offsets the reduction in resistance, the intended flow increase may not occur.

Distribution creates another difficulty. A vasodilator may act on vessels already carrying blood without reopening a channel blocked by a thrombus or cellular obstruction. Increasing flow through the accessible pathways does not guarantee restoration of the missing exchange surface.

Boerma and colleagues tested intravenous nitroglycerin against placebo after protocolized systemic resuscitation in severe sepsis or septic shock. The groups did not differ in the primary sublingual microvascular flow measure. The result shows that this regimen in this setting did not provide the proposed additional benefit. It does not establish that every vasodilator has identical effects in every patient. 10

Raising arterial pressure also requires a mechanism to justify it. The response depends on the downstream pressure, regional resistance, cardiac output, and local vessel behavior. Vasopressor escalation can raise pressure while increasing ventricular afterload or altering regional distribution. A higher arterial number alone does not establish improved exchange.

Treatment should address the supported limitation and include a planned reassessment. If the expected flow or perfusion benefit does not occur, reconsider the explanation. If congestion, arrhythmia, or impaired ventricular performance develops, reassess whether any observed benefit outweighs those effects.

15.11 Bringing Delivery and Exchange Together

Consider an illustrative patient with septic shock after initial resuscitation. Mean arterial pressure and cardiac output have improved. Capillary refill remains prolonged, lactate is elevated, and urine output is low. Arterial oxygen saturation is satisfactory, while central venous oxygen saturation is relatively high.

First separate what these observations show. Restored pressure and output suggest that the larger circulation has improved. Arterial saturation describes hemoglobin oxygenation, but does not establish an adequate hemoglobin concentration. The high central venous saturation describes blood returning through the sampled circulation; it does not certify oxygenation in every tissue region.

Uneven flow is one possible explanation for persistent abnormalities. Blood may return from well-perfused pathways with oxygen remaining while other regions receive too little. High delivery relative to demand or reduced cellular oxygen use can also raise venous saturation. None of these explanations can be selected from the saturation alone.

Confirm the measurements and examine their timing. Review hemoglobin, oxygenation, ventricular function, and rhythm. Assess venous pressure, cumulative fluid balance, and evidence of pulmonary or systemic congestion. Consider local obstruction when the organ findings suggest it. Infection treatment and source control remain essential because hemodynamic adjustments alone do not remove the cause of septic shock.

Repeat the peripheral examination with consistent technique and account for temperature and vascular tone. If sublingual imaging is used, confirm that the recording is technically adequate and represents more than one field. A poor-quality clip cannot establish the need for a new intervention.

Further fluid requires a reasonable expectation that increased flow will help and that the patient can tolerate the volume. Fluid responsiveness and fluid tolerance answer different questions. Rising stroke volume demonstrates a flow response; worsening congestion may still make additional fluid undesirable. Vasopressor escalation similarly requires a supported pressure limitation and reassessment of cardiac and regional effects.

Follow the direction of several findings over time. Refill can change before urine output recovers, and lactate may remain elevated for reasons beyond inadequate oxygen delivery. Persistent dysfunction should lead to a renewed explanation of its cause. It should not trigger indefinite escalation simply because one marker has not normalized.

References

1. Fåhraeus R, Lindqvist T. The viscosity of the blood in narrow capillary tubes. Am J Physiol. 1931;96(3):562–568. doi:10.1152/ajplegacy.1931.96.3.562.

2. Krogh A. The number and distribution of capillaries in muscles with calculations of the oxygen pressure head necessary for supplying the tissue. J Physiol. 1919;52(6):409–415. doi:10.1113/jphysiol.1919.sp001839.

3. Ellis CG, Bateman RM, Sharpe MD, Sibbald WJ, Gill R. Effect of a maldistribution of microvascular blood flow on capillary O2 extraction in sepsis. Am J Physiol Heart Circ Physiol. 2002;282(1):H156–H164. doi:10.1152/ajpheart.2002.282.1.H156.

4. De Backer D, Creteur J, Preiser JC, Dubois MJ, Vincent JL. Microvascular blood flow is altered in patients with sepsis. Am J Respir Crit Care Med. 2002;166(1):98–104. doi:10.1164/rccm.200109-016OC.

5. Sakr Y, Dubois MJ, De Backer D, Creteur J, Vincent JL. Persistent microcirculatory alterations are associated with organ failure and death in patients with septic shock. Crit Care Med. 2004;32(9):1825–1831. doi:10.1097/01.CCM.0000138558.16257.3F.

6. Ince C, Boerma EC, Cecconi M, et al. Second consensus on the assessment of sublingual microcirculation in critically ill patients: results from a task force of the European Society of Intensive Care Medicine. Intensive Care Med. 2018;44(3):281–299. doi:10.1007/s00134-018-5070-7.

7. Hernández G, Ospina-Tascón GA, Damiani LP, et al.; ANDROMEDA-SHOCK Investigators. Effect of a resuscitation strategy targeting peripheral perfusion status vs serum lactate levels on 28-day mortality among patients with septic shock: the ANDROMEDA-SHOCK randomized clinical trial. JAMA. 2019;321(7):654–664. doi:10.1001/jama.2019.0071.

8. Hernández G, Ospina-Tascón GA, Kattan E, et al.; ANDROMEDA-SHOCK-2 Investigators. Personalized hemodynamic resuscitation targeting capillary refill time in early septic shock: the ANDROMEDA-SHOCK-2 randomized clinical trial. JAMA. 2025;334(22):1988–1999. doi:10.1001/jama.2025.20402.

9. Ospina-Tascón G, Neves AP, Occhipinti G, et al. Effects of fluids on microvascular perfusion in patients with severe sepsis. Intensive Care Med. 2010;36(6):949–955. doi:10.1007/s00134-010-1843-3.

10. Boerma EC, Koopmans M, Konijn A, et al. Effects of nitroglycerin on sublingual microcirculatory blood flow in patients with severe sepsis/septic shock after a strict resuscitation protocol: a double-blind randomized placebo controlled trial. Crit Care Med. 2010;38(1):93–100. doi:10.1097/CCM.0b013e3181b02fc1.