Blood ﬂow, capillary transit times, and tissue oxygenation: the centennial of capillary recruitment

.—The transport of oxygen between blood and tissue is limited by blood’s capillary transit time, understood as the time available for diffusion exchange before blood returns to the heart. If all capillaries contribute equally to tissue oxygenation at all times, this physical limitation would render vasodilation and increased blood ﬂow insuf-ﬁcient means to meet increased metabolic demands in the heart, muscle, and other organs. In 1920, Danish physiologist August Krogh was awarded the Nobel Prize in Physiology or Medicine for his mathematical and quantitative, experimental demonstration of a solution to this conceptual problem: capillary recruitment, the active opening of previously closed capillaries to meet metabolic demands. Today, capillary recruitment is still mentioned in textbooks. When we suspect symptoms might represent hypoxia of a vascular origin, however, we search for relevant, ﬂow-limiting conditions in our patients and rarely ascribe hypoxia or hypoxemia to short capillary transit times. This review describes how natural changes in capillary transit-time heterogeneity (CTH) and capillary hematocrit (HCT) across open capillaries during blood ﬂow increases can account for a match of oxygen availability to metabolic demands in normal tissue. CTH and HCT depend on a number of factors: on blood properties, including plasma viscosity, the number, size, and deformability of blood cells, and blood cell interactions with capillary endothelium; on anatomical factors including glycocalyx, endothelial cells, basement membrane, and pericytes that affect the capillary diameter; and on any external compression. The review describes how risk factor-and disease-related changes in CTH and HCT interfere with ﬂow-metabolism coupling and tissue oxygenation and discusses whether such capillary dysfunction contributes to vascular disease pathology.


BACKGROUND
How does the vascular system provide each tissue cell with the oxygen supply needed for its functions and survival?This question is fundamental to our understanding of vascular physiology, tissue energy metabolism, and vascular disease pathology but not straightforward to answer.
Among his many contributions to science (47,55), Danish physiologist August Krogh pioneered the study of oxygen transport in tissue (34).Inspired by his supervisor Christian Bohr and working with his wife and closest collaborator, Marie Krogh, August Krogh applied mathematical models, developed unique precision instruments, and carried out numerous animal studies to demonstrate the importance of free diffusion for oxygen transport in tissue (34).Importantly, August Krogh realized that the oxygenation benefits of higher blood supply, which we still take for granted, are counteracted by the shorter time blood remains in the capillaries (34).This review describes how this realization, and Krogh's Prize-winning studies to understand how oxygen transport could still be effective as blood flow increases, have inspired a century of research into the roles of capillaries in tissue oxygenation and vascular disease pathology.
Although the number of capillaries in the human body is astronomical, and their topology varies from one organ to another, tissue oxygenation has historically been inferred from the properties of a single capillary (53).Figure 1A illustrates how bloodtissue oxygen transport depends on blood's capillary flow velocity.While the supply of oxygenated blood increases linearly with erythrocyte speed, the time spent by each erythrocyte within the capillary falls rapidly.Given as the ratio between capillary volume and is blood flow, the capillary transit time determines the time available for blood-tissue diffusion exchange before blood leaves the tissue.Because of these opposing effects, blood-tissue oxygen transport is efficient for moderate capillary erythrocyte velocities, but at higher blood velocities, further velocity increases do little to improve oxygen availability in tissue outside.
The relation in Fig. 1 originated with Christian Bohr, August Krogh's mentor and laboratory director, who developed the formula to estimate lungs' capacity to transport oxygen into the circulation by passive diffusion (9).Bohr's formula was later developed independently by other physiologists and now referred to as the Bohr-Kety-Crone-Renkin equation or simply the flow-diffusion equation (53).Assuming that all capillaries have identical extraction properties, the equation predicts the extraction of freely diffusible molecules in tissue, based on its blood flow F, the surface area of its capillaries S, and the capillary wall's permeability P to the molecules (53) (Fig. 1B).
Few may recollect the flow-diffusion equation, but its importance for our conceptualization of vascular function cannot be overstated.This owes partly to its fundamental predictions regarding tissue oxygenation, which we all take for granted: Property I: tissue oxygenation is determined by its blood flow and capillary surface area.Property II: tissue oxygenation improves as its blood flow increases.Bohr (9) and Krogh (29) both realized that Bohr's equation could not account for the lungs' oxygen uptake during exercise by diffusion, even with Krogh's precise measurements of oxygen tensions in lung alveoli and blood.Bohr therefore remained convinced that some alveolar oxygen is actively secreted into the bloodstream, while August and Marie Krogh correctly surmised that lung oxygen exchange is driven by passive diffusion (29).

AUGUST KROGH'S CAPILLARY RECRUITMENT HYPOTHESIS
Soon after dismissing his mentor's concerns, August Krogh faced a similar problem as he tried to account for muscle oxygen uptake during exercise.For simplicity, he had approximated muscle tissue by parallel, identical tissue cylinders supplied by a central capillary ("Krogh cylinders") and asked mathematician Agner K. Erlang to calculate the oxygen tension throughout these cylinders, assuming that tissue consumes oxygen at a certain rate as it diffuses away from the capillary (28).Using this Krogh-Erlang equation with data available at the time, however, he was unable to explain observations he had made in human volunteers a few years earlier, namely, that a 10-fold increase in muscle blood flow during ergometer-training could be supported by a 10-fold increase in muscle blood flow (28).Guided by the Krogh-Erlang equation, he arrived at the surprising hypothesis that most muscle capillaries (90%) in fact remain closed during rest, but open during exercise, so-called capillary recruitment (31).Krogh's proposal is consistent with the flow-diffusion equation's predictions (Fig. 1B): By increasing the capillary surface area 10-fold, the capillary bed's oxygen transport capacity increases in parallel.In essence, Krogh's capillary recruitment hypothesis offered a biological mechanism to mitigate transit-time effects: as blood flow and the number of open capillaries (capillary blood volume) increase in parallel, capillary transit times, and thereby oxygen extraction efficacy, remain constant.
Having formulated his hypothesis, Krogh quantified oxygen diffusion in tissue (30), counted the number of open capillaries in muscle cross section from animals he had injected with ink, and observed the opening and closing of capillaries in vivo (28,31).Combining his data with the Krogh-Erlang equation, he concluded that capillary recruitment can account for muscle oxygen uptake during rest and exercise (31).Having published his Fig. 1.A: the amount of oxygen that tissue can extract from a capillary increases substantially as erythrocyte flow-through increases, particularly for slow or modest erythrocyte velocities.Note, however, how doubling erythrocyte velocity has a less-than doubling effect on oxygen availability at higher velocities.B: the corresponding relation for a tissue volume with a large number of capillaries with equal length and blood flow.The flow-diffusion equation for a tissue volume relates blood flow (F) to tissue oxygenation (O), defined as the steady-state oxygen uptake that blood flow can support at normal tissue oxygen tension, with C a denoting arterial oxygen concentration, P the capillary wall permeability to oxygen, and S the capillary surface per unit volume.Note that the curve approaches a theoretical maximum, C a •PS, as blood flow becomes very high.Also note that both capillary dilation and opening of previously closed capillaries cause S to increase, improving oxygenation independent of blood flow.Representations of the capillary bed as parallel capillaries of equal length, in this case also equal flows, helps conceptualize capillary function, e.g., in muscle tissue.Across organs, the orientation, length, and interconnectedness of capillary segments vary greatly as they adapt to the microscopic organization of cells and their metabolism (50) and oxygen transport becomes more complex to model (16).
observations in 1919 (28,30,31), he was awarded the Nobel Prize for "his discovery of the capillary motor regulating mechanism" the following year.

THE FATE OF CAPILLARY RECRUITMENT
The concept of capillary recruitment gained widespread acceptance after August Krogh's Nobel Prize and is still widely mentioned in text-books and scientific literature, often to account for the additional "capillary surface area" needed to explain the transport of oxygen and other diffusible molecules into tissue during activation.Modern studies, however, find too few closed capillaries for capillary recruitment to account for tissue oxygenation in muscle and brain (33,46), and concerns have been raised that Krogh's observations of closed capillaries may have been linked to inherent limitations of experimental approaches available at the time (48).
Krogh speculated that Rouget cells, now known as pericytes, possess the contractile properties required by his "capillary motor mechanism" (31).Subsequent studies from his laboratory supported this notion, but rather than opening and closing capillaries, these observations suggested that pericyte activity cause more modest changes in capillary diameter under physiological conditions (8).Pericytes serve multiple functions (5) and their phenotype seemingly depends on species, tissue type, and their location along the arteriole-venule axis (61).This complexity, and changing naming conventions for contractile cells at the arteriole-capillary transition (7,17,22), have so far conspired against a consensus on the active control of capillary diameter across tissue types.

TRANSIT-TIME HOMOGENIZATION
With capillary recruitment an unlikely mechanism, blood-tissue oxygen transport remained unaccounted for.Decades later, indicator-dilution experiments began to question whether capillary transit-times are uniform under resting conditions (54), invalidating the extrapolation from single capillary properties to those of tissue made by Bohr, Krogh, and others (53).
During indicator-dilution experiments, diffusible molecules and an intravascular reference substance are bolus injected into the arterial inflow of an organ and their venous outflow concentrations monitored.From these data, blood-tissue exchange of the diffusible molecules, blood flow, and the distribution of vascular transit times through the organ can be characterized.In the heart, such studies suggested that capillary transit times are very heterogeneous during rest but homogenize as coronary resistance vessels dilate, and only then could the Krogh-Erlang equation account for the extraction of diffusible molecules (54).
Direct in vivo studies of the brain's microcirculation confirmed that erythrocyte flows vary greatly across the capillary bed during rest but homogenize as blood flow and oxygen demands increase (56,62).In muscle tissue, this homogenization is accompanied by an increase in HCT, further increasing the number of oxygen-carrying erythrocytes across capillaries at any given time (46).Note that higher HCT increases the oxygen concentration in capillary blood and represents an instance of prolonged capillary transit times for cells (erythrocytes) only.

EXTENDING THE KROGH-ERLANG AND FLOW-DIFFUSION EQUATIONS
These observations lead to the introduction of flow heterogeneity into biophysical models, and the single-capillary flow-diffusion (Fig. 1A) and Krogh-Erlang equations have now been generalized to tissue while taking capillary transit-time heterogeneity (CTH) into account.In addition to tissue blood flow or blood mean transit time (MTT; the ratio between capillary blood volume and blood flow), these general equations therefore include the standard deviation of capillary blood flows or transit times (2,3,26,52).
Figure 2 illustrates how homogenization of capillary flows across the capillary bed improves oxygen extraction as blood flow increases, in spite of the single-capillary properties shown in Fig. 1A.
Inserting observed changes in blood flow, CTH, and HCT during exercise, the general Krogh-Erlang equation can now account for the 80-fold increases in muscle oxygen uptake measured during exercise without capillary recruitment (2), while the general flow-diffusion equation can account for oxygen uptake in brain tissue during functional activation, based on parallel blood flow and CTH changes (52).

IS CAPILLARY TRANSIT TIME HOMOGENIZATION AN ACTIVE PROCESS?
The microcirculation's topology and morphology optimize the distribution of erythrocytes to meet cellular metabolic demands (39,50).Simulations of blood flow through microvascular networks suggest that the standard deviation of blood's transit-times, CTH, decreases in proportion to their mean, MTT, as blood flow increases through passive, compliant microvascular networks (52).Indeed, in rodent brain, moderate hypotension and carbon dioxide (CO 2 ) inhalation are accompanied by parallel changes in MTT and CTH (19,24,26).In rodent brain, however, CTH drops more than MTT during brain activation (19) and capillary flow homogenization seemingly occurs before the increase in blood flow (36).Such active contributions to capillary flow homogenization in relation to increased blood flow may include increased erythrocyte deformability in response to lower tissue oxygen levels (63), active regulation of capillary diameter by pericytes in response to local cell signaling (20), or other signals associated with oxygen supply-demand imbalance, such as reduced tissue oxygen levels (32) in conjunction with increasing lactate levels (64).

CAPILLARY DYSFUNCTION: UNEXPECTED CONSEQUENCES OF CAPILLARY FLOW DISTURBANCES
The flow-diffusion equation, first formulated by Christian Bohr, fails to explain oxygen extraction at high flow due to its assumption that all capillary flows are equal.August Krogh's recruitment hypothesis overcame this shortcoming by proposing that the number of equally perfused capillaries increases with blood flow to meet metabolic demands.In healthy tissue, Krogh's model shares similarities with our current view of capillaries' role in tissue oxygenation as he merely assumed that a binary, rather than a continuous, distribution of capillary flows undergoes homogenization as blood flow increases (2).In disease, however, neglecting flow differences among capillaries has profound, conceptual implications, as it limits us to think of tissue hypoxia and vascular pathology in terms of conditions that limit blood supply and cause ischemia.In fact, due to the nonlinear relation between capillary flow velocity and oxygen extraction shown in Fig. 1A, flow changes in individual capillaries can, biophysically, limit tissue oxygenation to the same extent as reductions in blood flow (70).
Figure 3 shows how our intuitive understanding of tissue oxygenation, and even properties I and II above, break down if blood flow and its capillary distribution no longer change in parallel.Capillary dysfunction denotes a gradual increase in resting CTH and the failure of CTH to decrease during episodes of increased blood flow.Figure 2 illustrates why such changes, if unmitigated, reduce oxygenation at any blood flow and limit the increase in oxygen metabolism that can be supported by increased blood flow.
Capillary dysfunction may be the result of accumulating damage to the constituents of the capillary wall (the glycocalyx, endothelial cells, basal membrane, and pericytes) due to aging, risk factors, or disease, but may also be the result of altered blood properties (e.g., plasma viscosity, erythrocyte deformability, and endothelial leukocyte or erythrocyte adhesion) or external compression (e.g., edema, elevated intracranial pressure).
Figure 4A illustrates such capillary flow disturbances.Note that "microischemia" (27) may threaten the energy metabolism of individual cells if most nearby capillaries become occluded or poorly perfused.The greater problem, however, is the disproportionate amount of blood that passes through other capillaries at short transit times, introducing "shunts" for oxygenated blood through the tissue.
Tissue's oxygen supply-demand balance is determined by its blood flow, CTH, and capillary HCT on one hand, and its oxygen utilization, on the other.In most tissues, blood flow is regulated to maintain tissue oxygen tension (P t O 2 ) within a narrow range.Should capillary dysfunction cause oxygen extraction to fall slightly, blood flow is therefore expected to increase to maintain constant P t O 2 .At normal P t O 2 , blood is typically more than half-saturated with oxygen as it leaves the capillary bed, in near-equilibrium with tissue.The corresponding oxygen extraction fraction (OEF; typically 30-50%) increases if blood flow falls, at the expense of a drop in P t O 2 .
Figure 4C summarizes the blood flow changes required to preserve P t O 2 and tissue oxygen utilization if worsening capillary dysfunction causes CTH to increase gradually.Also shown are parallel reductions in P t O 2 .These are expected to activate hypoxia-sensing mechanisms, inflammation (13), and energysaving adaptations that ultimately affect tissue function.In brain tissue, hypoxia also stimulates the formation and retention of neurotoxic proteins (57) that, in turn, constrict pericytes (44).

DO WE OBSERVE DISEASE PATTERNS CONSISTENT WITH CAPILLARY DYSFUNCTION?
The earliest sign of capillary dysfunction is expected to be a compensatory blood flow increase: the exact opposite of what we now expect of vascular (flow-limiting) disease changes (Fig. 4C).In the absence of microischemic damage, this phase of capillary dysfunction is not expected to affect cell functions because Fig. 3. A: the Chesler counterexample, named after the contributing coauthor (73), demonstrates that tissue oxygenation decreases with any capillary flow heterogeneity, independent of blood flow and capillary surface area.The example disproves the common assumption that tissue oxygenation can be gleaned from its blood flow alone (property I in text): imagine tissue with n identically perfused capillaries.Leukocytosis slows blood velocities through half of the capillaries while blood velocities increase through the remaining to maintain blood flow.Points labeled A and B represent identical blood flow but different tissue oxygenations.B: the Jespersen counterexample, named after the contributing author (26), disproves the widely held assumption that increased blood supply inevitably improves tissue oxygenation (property II in text): From the same initial condition, imagine blood flow increases before leukocytosis slows blood velocities through half of the n capillaries.Blood flow increases, but oxygenation declines, between points labeled A and B. overall P t O 2 is preserved.Presymptomatic hyperemia is observed in several conditions that affect the microcirculation, including Alzheimer's disease (AD): carriers of the APOE-ɛ4 gene allele, which increases the risk of developing late-onset AD severalfold, show areas of elevated cerebral blood flow (CBF) decades before some develop dementia (66).These findings lead some to rule out a vascular etiology of AD but could in fact signify a role for capillary dysfunction in the disease.
Using indicator-dilution techniques and MRI (40), microvascular flow disturbances can now be observed in AD patients (15) (Fig. 5).The severity and 6-mo progression of AD patients' cognitive symptoms, respectively, correlate with regional microvascular flow disturbances (42) and the onset of mild cognitive impairment (MCI) seemly coincides with the transition from hyperemia to limited blood flow across patients' brains (43).
In the absence of intrinsic, compensatory flow attenuation, the progression of capillary flow disturbances is expected to gradually reduce P t O 2 (Fig. 4B).In some types of cerebral small vessel disease, infarction is in fact preceded by normal or increased blood flow (67).

BLOOD PROPERTIES AS RISK FACTORS
Transit time effects may help conceptualize how blood properties interfere with tissue oxygenation.Children with sickle cell anemia (SCA), for example, are at extreme risk of stroke and brain small-vessel vasculopathy, especially those with high cerebral blood flow (CBF) as measured by transcranial Doppler (58).Hemoglobin S has reduced oxygen affinity and polymerizes upon deoxygenation, causing erythrocytes to become more rigid and adhere to venular endothelium.The conformational changes reduce erythrocyte oxygen carrying capacity and deformability, and their premature lysis causes anemia with patient systemic hematocrits 2/3 of normal values (58).Notably, transfusions with normal blood dramatically reduce the risk of stroke in patients with high CBF (58).Cerebral microvessels are seemingly maximally dilated in SCA (21), maximizing blood's capillary transit times despite the high flow.As illustrated by Fig. 2, tissue oxygenation is expected to be far more sensitive to capillary flow disturbances (higher CTH) than to reductions in blood flow under high-flow conditions.Capillary flow disturbances are therefore likely contributors to hypoxic tissue injury in these patients, even in the absence of flow-limiting conditions.
So far, the capillary distribution of oxygen has been inferred from the distribution of erythrocytes in terms of their HCT and velocity.Erythrocyte sizes vary slightly, with unexpectedly low and high hemoglobin content in the smallest and largest erythrocytes, respectively, considering their volumes (38).Microscopic tissue oxygenation is an extremely complex function of the distribution of red blood cells and their oxygen content across the capillary network (23,51).Should larger, oxygen-rich erythrocytes show any preference for capillary "shunting" (cf.Fig. 4A), however, higher red blood cell distribution width (RDW) would be expected to worsen the oxygenation effects of capillary dysfunction.Such effects could add to the unexpected, graded relations between RDW, morbidity, and mortality found across a range of conditions that affect the microcirculation (60).

THERAPEUTIC IMPLICATIONS?
In conditions with severe capillary dysfunction, augmented blood flow may not improve tissue oxygenation, violating property I above.If large vessel disease with ischemic symptoms is accompanied by capillary changes, the latter may therefore limit the benefits of blood flow normalization.Severe carotid artery Fig. 4. A: changes in capillary patency can disturb capillary flow patterns and prevent their normal homogenization during increases in blood flow.B: if these disturbances accumulate, the gradual increase in capillary transit times (CTH) is expected to limit oxygen uptake, causing tissue oxygen tension (P t O 2 ) to decrease in the absence of compensatory blood flow changes.C: compensatory hyperemia may maintain normal P t O 2 for mild capillary dysfunction, but for severe capillary dysfunction, the "shunting" of oxygenated blood can only be mitigated by limiting blood flow, prolonging blood transit times to increase oxygen extraction fraction.Such compensatory flow changes may delay symptoms and tissue damage, as illustrated by the time at which P t O 2 starts to drop and reaches 50% of normal values, respectively.stenosis is indeed accompanied by downstream capillary flow disturbances (6) that partly account for regionally impaired brain oxygenation (increased OEF) (69).Carotid revascularization, however, seems to normalize both blood flow and capillary flow patterns in the affected tissue (6) (Fig. 6).
Animal studies suggest that brain and heart pericytes constrict during prolonged ischemia, possibly hindering blood reflow after recanalization (20,45,65).Even if blood flow is restored, persisting capillary flow disturbances would therefore be expected to maintain tissue hypoxia and could cause reperfusion injury (26,71).Early reperfusion seems to reverse microvascular flow disturbances that develop during the ischemic period in acute stroke patients (14).In acute stroke patients recanalized later after symptom onset by thrombectomy, however, capillary dysfunction could contribute to the lack of clinical improvement observed in more than 30% of cases, so-called futile reperfusion (37).In hypoperfused brain tissue where reperfusion cannot be achieved, however, tissue survival seemingly depends on the level of microvascular flow disturbances, independent of residual blood flow (14).This may imply that early protection of microvascular function can salvage brain function, even in patients whose blood flow cannot subsequently be restored.
Early vascular disease phenomena, such as hypertension and neurovascular/endothelial dysfunction, involve changes in vascular smooth muscle cell (VSMC) tone and blood flow responses (25) that resemble those changes predicted to mitigate capillary dysfunction (Fig. 4B).Whether or not increased vascular resistance in certain organs serves to optimize tissue oxygenation as capillary dysfunction reaches critical levels, it may prove fruitful to consider capillary function as a therapeutic target: Thus, while the restoration of normal capillary flow homogenization, biophysically, always improves tissue oxygenation (Figs. 2 and 3A), this is not universally true for therapeutic "normalization" of VSMC tone and blood flow responses (Fig. 3B).To this end, VSMCs and downstream pericytes differentially express receptors to the same vasoactive molecules (12).Meanwhile, tools to assess key blood properties and to routinely image capillary function in key organs (cf.Fig. 5) may help future, individualized monitoring and management of capillary health.
For drug development in general, it may prove important for animal models to reflect the level of capillary dysfunction typical of the corresponding human condition.In the case of AD and ischemic stroke, for example, residual capillary dysfunction and tissue hypoxia after therapeutic removal of neurotoxic proteins or arterial occlusions, respectively, may differ between patients and animal models and unexpectedly limit translational success.

TRANSIT TIME EFFECTS IN CRITICAL ILLNESS
Capillary dysfunction may play a role in severe sepsis, which seemingly violates property I as organ failure and hypoxic tissue injury progresses despite adequate systemic oxygen delivery.Capillary flow disturbances, caused for example by breakdown of the capillary glycocalyx or leukocyte adhesion to capillary endothelium, may be the source of severe hypoxia in the heart, kidney and brain, despite preserved blood flow (68).Microthrombosis with occlusion of individual capillaries is a common finding in critically ill patients, SCA patients with vaso-occlusive crisis (VOC) (58) and more recently, COVID-19.Due to the parallel, interconnected nature of microvascular networks, the closing of individual capillary segments may not limit organ blood flow.The redistribution of blood to less affected capillary pathways, however, shortens and disturbs capillary transit times and may therefore be a source of hypoxic tissue injury and inflammation.

TRANSIT TIME EFFECTS IN THE LUNG CIRCULATION
Analogous to oxygen uptake in tissue, oxygen uptake in the lungs is limited by short capillary transit times, and transit time homogenization seemingly counteracts hypoxemia when cardiac output is high and alveolar capillary transit times therefore short (49).Alveolar oxygen transport models corroborate that a homogenous distribution of erythrocytes across alveolar capillaries contributes significantly to the efficient alveolar oxygen uptake in lungs ( 23) that once puzzled Bohr and Krogh.To further optimize blood oxygenation across the lungs, hypoxic vasoconstriction limits blood flow to poorly ventilated alveoli, matching alveolar ventilation and perfusion.This mechanism is particularly important in acute respiratory distress syndrome, as it limits blood flow through fluid-filled alveoli.
Some SCA patients develop severe hypoxemia 2-3 days after VOC.Analogous to oxygen uptake in the brain, blood oxygen uptake in the lung is extremely sensitive to capillary flow disturbances if blood flow through the lungs is high.If cardiac output is elevated to prevent tissue hypoxia in other organs, increases in alveolar capillary CTH may therefore cause hypoxemia.Even patients with limited signs of perfusion-ventilation mismatch on diagnostic imaging may therefore be severely hypoxemic due to rigid and adherent erythrocytes, capillary thrombosis, and granulocyte recruitment due to infection or inflammation (58).Notably, due to CO 2 's high diffusivity compared with oxygen, hypercapnia is not expected to develop in such "isolated" alveolar capillary dysfunction.Such "silent hypoxia" was recently linked to COVID-19 infection and may be related to embolization of a significant proportion of alveolar capillaries (1).

CONCLUSION
August Krogh had a unique ability to identify scientific problems that, when examined further, generated scientific breakthroughs and establish new, fertile research areas (55).Krogh himself was quick to dismiss his own hypotheses if they failed to stand the test of experiments (55), and in many ways, his very realization that tissue oxygenation depends on both blood flow and capillary function may prove as important to the advancement of science as his capillary recruitment hypothesis and the notion that capillary diameters are actively regulated.
Have we overlooked capillary dysfunction as a "silent" source of age-and disease-related tissue hypoxia by linking vascular disease to the more conspicuous blood flow limitations?To address this question further, new diagnostic methods and biomarkers must be developed to allow easy assessment of capillary function and tissue oxygenation across organs (72).Thus far, findings support the hypothesis that capillary transit time effects play a role in vascular disease pathology (4,6,11,14,41), as well as in conditions less associated with vascular disease pathology (35), including AD (15,18,42,43) and cancer (10,59).

Fig. 2 .
Fig. 2. The original flow-diffusion equation (full line) assumes identical capillary transit times (CTH = 0) while the extended flow-diffusion equation predicts the relation between tissue oxygenation and blood flow for different levels of CTH (dashed line).The lower and upper capillary beds illustrate relative capillary blood velocities and saturations (red = 100%, blue = 65%) in resting and activated brain tissue, respectively (26).Note how oxygen availability at resting blood flow is much lower than previously thought due to high CTH.The combination of increased blood flow and reduced CTH allows (red arrow) tissue oxygenation to increase linearly with blood flow.Note how higher CTH (blue arrow) always reduces oxygenation at a given blood flow.

Fig. 5 .
Fig. 5. Parametric maps of the relative transit time heterogeneity (RTH) in 2 healthy controls (top row) and in patients with mild cognitive impairment (MCI) and Alzheimer's disease (AD), respectively (bottom row, left to right).[Reprinted with permission from Elsevier(15).]Maps were calculated from dynamic susceptibility contrast MRI data(40).Due to the tendency for capillary transit times and mean transit time to covary, their ratio, RTH, is a convenient index of microvascular flow disturbances.MMSE, Mini-Mental State Examination score.