Best Practice & Research Clinical Anaesthesiology 28 (2014) 441e451

Contents lists available at ScienceDirect

Best Practice & Research Clinical Anaesthesiology journal homepage:


Monitoring the microcirculation in critically ill patients Daniel De Backer, Professor *, Arthur Durand, Fellow Department of Intensive Care, Erasme Hospital, Universit e Libre de Bruxelles, Brussels, Belgium

Keywords: tissue perfusion tissue oxygenation videomicroscopy tissue PCO2 laser Doppler

Alterations in microvascular perfusion have been identified in critically ill patients, especially in sepsis but also in cardiogenic shock, after cardiac arrest, and in high-risk surgery patients. These alterations seem to be implicated in the development of organ dysfunction and are associated with outcome. Even though microvascular perfusion can sometimes be homogenously decreased as in acute hemorrhage or in non-resuscitated cardiogenic shock, heterogeneity of perfusion is observed in sepsis and in resuscitated hemorrhagic/cardiogenic shock. Heterogeneity of perfusion has major implications for monitoring, as many techniques cannot detect microcirculatory alterations when heterogeneity of flow is present in significant amount. Indeed, devices such as laser Doppler or O2 electrodes and near-infrared spectroscopy have a relatively large sampling volume and measurements are affected by the highest values in the field. Using these techniques during a vascular occlusion test may help to characterize microvascular reactivity; however, microvascular reactivity sometimes fails to represent actual microvascular perfusion. Videomicroscopic techniques can nowadays be applied at bedside but are still restricted to some selected patients (quiet or sedated patients). Tissue PCO2 is an elegant alternative but is not yet broadly used. In this manuscript, we discuss the main advantages and limitations of the techniques available for bedside evaluation of the microcirculation in critically ill patients. © 2014 Elsevier Ltd. All rights reserved.

* Corresponding author. Department of Intensive Care, Erasme University Hospital, Route de Lennik 808, B-1070, Brussels, Belgium. Tel.: þ32 2 555 3380; Fax: þ32 2 555 4698. E-mail address: [email protected] (D. De Backer). 1521-6896/© 2014 Elsevier Ltd. All rights reserved.


D. De Backer, A. Durand / Best Practice & Research Clinical Anaesthesiology 28 (2014) 441e451

Introduction Shock is characterized by an impairment in tissue perfusion, of various causes, leading to impaired metabolism and organ dysfunction, and associated poor outcome [1]. Even though there is no doubt that initial therapy should aim at achieving some minimal values of arterial pressure and cardiac output, further increasing blood pressure and/or cardiac output, once these initial goals are reached, often fails to improve tissue oxygenation and outcome [2]. There are basically three levels in the journey of oxygen from the heart to the cells. The first level is the systemic circulation, constituted by cardiac output, oxygen content, and blood pressure. The second level is at the regional level, mostly constituted by the microcirculation, which is responsible for the distribution of flow inside each organ. The third level is at the mitochondrial level, which constitutes cytochrome chain and responsible for oxygen utilization. Recent trials have shown that goal-directed therapy aiming at optimizing systemic oxygen transport failed to improve survival [3,4]. This suggests that alterations either in microvascular perfusion or in oxygen utilization were not improved by optimization of systemic hemodynamics. Alterations in microvascular perfusion frequently occur in critically ill patients, and especially in patients with sepsis [5]. As these often persist after the correction of systemic hemodynamic abnormalities [5,6], and as their severity is associated with a poor outcome [5e9], it is important to raise awareness on the characteristics of the diseased microcirculation and on the tools available to evaluate the microcirculation “at the bedside”. Different techniques can be used to evaluate microcirculation at bedside. The choice of the device or technique should be guided by the expected type of alteration, which mostly depends on the underlying disease (i.e., sepsis, trauma). In this manuscript, we discuss the techniques that are the most currently used at the bedside to assess the microcirculation in critically ill patients. Microcirculation alterations in critically ill patients Normal microcirculation is characterized by a dense network of perfused capillaries (Fig. 1). In normal conditions, there is minimal heterogeneity, most of the visualized capillaries being perfused even though flow in the various capillaries varies according to metabolic needs of the surrounding tissues. Adaptation to metabolic needs occurs by opening and closing capillaries, and adapting the velocity of circulating cells within these. Modulation of precapillary sphincters is partly under the influence of systemic factors, with sympathetic stimulation and circulating substances but most of

Fig. 1. Sublingual microcirculation of a healthy individual. Visualization of sublingual microcirculation of a healthy individual. Note the rich density of capillaries, all of which are perfused.

D. De Backer, A. Durand / Best Practice & Research Clinical Anaesthesiology 28 (2014) 441e451


the fine-tuning perfusion is regulated by local factors that include direct stimulation of endothelial cells by backward communication and local release of nitric oxide by red blood cells under hypoxic conditions. In sepsis, capillary density is decreased and a variable proportion of capillaries is not perfused or at best intermittently perfused, leading to heterogeneous tissue perfusion [10e12] (Fig. 2). These alterations have been reported in different models of sepsis and in all animal species [10,12e14], and observed in all studied organs, including the brain [13,14]. This process is not fixed, as capillary perfusion may change minute by minute, and non-perfused capillaries suddenly become perfused and vice versa. Despite the fact that sepsis-associated alterations in coagulation seem to be implicated in microcirculatory alterations as fibrin deposition occurs in non-perfused capillaries [12] and that anticoagulant agents can improve the microcirculation [15], these observations of transient impairment of capillary perfusion suggest that thrombosis does not play a major role. The dysregulation of communication at the endothelial level seems to be a key factor in these alterations. When submitted to hypovolemia, the normal microcirculation decreases its minimal heterogeneity in order to match microvascular oxygen delivery to oxygen consumption while the heterogeneity further increases in the septic microcirculation leading to mismatch between flow and oxygen requirements [16]. Similar events were demonstrated in septic patients. Compared to non-infected critically ill patients and healthy volunteers, the sublingual microcirculation of patients with severe sepsis is characterized by a significant decrease in vessels' density and in the proportion of perfused small vessels related to an increase in non-perfused as well as in intermittently perfused vessels [5]. In addition to being increased inside the visualized area, heterogeneity of perfusion is also increased between areas close by a few microns. Many groups confirmed these data [7,17,18]. Several studies illustrate the link between microvascular perfusion and organ dysfunction [6,7,9,18,19]. Microcirculatory alterations can be observed very early in the course of sepsis [18] but did not occur in patients with infection without signs of severity [20], suggesting that microcirculatory alterations contribute to the development of organ dysfunction. Evolution over time and/or in response to therapy is also very informative on this implication. Early improvements in microvascular perfusion occurring in response to goal resuscitation therapies are associated with subsequent improvements in organ function while organ function deteriorated when microvascular perfusion failed to improve or even deteriorated [9,19]. Microcirculatory alterations are more severe in non-survivors than in survivors [6,7]. In 252 septic patients, survival rate progressively decreased with quartiles of severity in

Fig. 2. Sublingual microcirculation of a septic patient. Visualization of sublingual microcirculation of a patient with septic shock. Capillary density is decreased compared to Fig. 1. In addition, several capillaries are not perfused (the arrow indicates a stopped flow capillary).


D. De Backer, A. Durand / Best Practice & Research Clinical Anaesthesiology 28 (2014) 441e451

alteration of the microcirculation [6]. In addition, alterations in microvascular perfusion were one of the strongest predictors of outcome and remained independently associated with outcome in multivariate analysis. Finally, the time course of microvascular alterations also differs between survivors and non-survivors. Microvascular alterations improved over time in response to therapy in survivors but not in non-survivors [7]. Microvascular alterations have also been observed in critically ill patients in many other conditions than sepsis: these have been reported in cardiogenic shock [21], after resuscitation from traumatic injury [22], and in patients after high-risk surgery [23]. These interventions were similar to those observed in sepsis, but often less severe. Nevertheless, these were often associated with organ dysfunction and poor outcome [21]. Sometimes our interventions also can induce microvascular alterations, and in particular anesthesia [24] and/or sedation [25]. Several interventions may affect the microcirculation. Interventions aimed at improving global hemodynamics may also have microvascular effects but these were independent from the changes in global hemodynamics [26,27]. Nevertheless, one should bear in mind that microvascular alterations are heterogeneous and that all interventions should aim at recruiting the microcirculation more than increasing flow in already perfused vessels. Fluids can improve microvascular perfusion, increasing the proportion of perfused capillaries and decreasing perfusion heterogeneity [26,28,29]. Interestingly, fluids improved the microcirculation only when administered within the first 24 h of sepsis and only the first fluid bolus was associated to microvascular perfusion improvement [28]. In septic patients identified as fluid respondents in terms of systemic circulation, organ dysfunction improved only in the patients who improved their microcirculation in response to fluids. The effects of red blood cell transfusions are variable and may depend on the severity of underlying microcirculatory alterations, with an improvement in microvascular perfusion in the most severe septic patients and a worsening in patients with microcirculation closer to normal values [30]. Vasoactive agents also affect the microcirculation. b-Adrenergic agents improve microvascular perfusion, even though these effects can be quite variable among individuals [27,31,32]. Other inotropic agents seem to also improve microvascular perfusion, but these were less well studied [33,34]. Vasopressor agents, when used to reverse severe hypotension, may improve microvascular perfusion [35,36], but increasing mean arterial pressures to higher levels had variable effects on microvascular perfusion [37,38]. Is there a place for vasodilatory agents? In a pilot trial, nitroglycerin administration was shown to improve the microcirculation [39] but these effects were not confirmed in a randomized trial [40]. Angiotensin-converting enzyme inhibition failed to improve the microcirculation in sepsis [41] but was beneficial in severe heart failure [42]. The usefulness of vasodilatory agents is thus far to be demonstrated. An attractive pathway may be modulation of endothelial NO synthase (eNOS), which was found to improve microvascular perfusion and organ function in experimental sepsis [43,44]. Techniques to investigate the microcirculation Microvascular perfusion can be measured directly by various methods or evaluated indirectly by indices of tissue oxygenation [45]. Importantly, all techniques evaluate the microcirculation in a given window and the capacity of this window to be representative of other vascular beds depends on the mechanisms implicated in the microcirculatory dysfunction (global decrease in perfusion such as in low flow conditions, systemic factors affecting all organs such as in sepsis, focal, and affecting mostly some organs such as in compartment syndrome or ischemia/reperfusion injury after organ transplantation), on organ microvascular architecture (i.e., gut or kidneys), and on local factors (such as local vasoconstriction, temperature, and pressure). The main characteristics of the various techniques and their potential usefulness are listed in Tables 1 and 2. Clinical evaluation and biomarkers An impaired microvascular perfusion may be inferred from a general clinical examination [46,47]. However, the skin is very sensitive to the effects of temperature and vasoactive agents, making the link

D. De Backer, A. Durand / Best Practice & Research Clinical Anaesthesiology 28 (2014) 441e451


Table 1 Characteristics of microcirculatory alterations in different settings. Disease



Device to investigate the microcirculation

Arteriolar constriction Low flow Venous constrictionþþþ

Hypoxia in all areas Low SvO2

Idem þ heterogeneity

Patchy hypoxic areas SvO2 variable

Laser Doppler Videomicroscopy Tissue PCO2 VOT Videomicroscopy Tissue PCO2 VOT

Arteriolar constriction Low flow Venous congestion

Hypoxia in all areas Low SvO2

- Reperfusion (resuscitated stage)

Idem þ heterogeneity

Patchy hypoxic areas SvO2 variable


Excessive flow in some capillaries þ stop flow in other capillaries

Patchy hypoxic areas High SvO2

Hemorrhage - Early

- Reperfusion (resuscitated stage) Cardiogenic - Early

Laser Doppler NIRS SO2 Videomicroscopy Tissue PCO2 VOT Videomicroscopy Tissue PCO2 VOT Videomicroscopy Tissue PCO2 VOT

NIRS SO2: measurement for O2 saturation with near-infrared spectroscopy. VOT ¼ vascular occlusion test assisted either with laser Doppler or NIRS and evaluating microvascular reactivity.

between skin hypoperfusion and more central beds quite loose [46]. Even though microvascular alterations are associated with increased lactate levels [20], lactate levels have a poor sensitivity and specificity to detect microcirculatory alterations [6,9,26,27]. Interestingly, changes in lactate levels after therapeutic interventions are proportional to the improvement in microvascular perfusion [27].

Table 2 Tools available to investigate the microcirculation in humans. Technique

What is measured?



Skin mottling/CRT

Skin regional and microvascular perfusion

Easy and cheap


Balance between O2 consumption and supply


Balance between O2 consumption and O2 requirements Velocity of red blood cells

Easy, minimally invasive, widely available, numeric targets Easy, cheap, widely available, numeric targets

Influenced by temperature and vasoactive agents, absence of numeric targets Affected by microvascular shunt, normal values can be misleading Delayed clearance, not always hypoxic origin

Laser Doppler

Easy, not expensive


Visual evaluation of microvascular perfusion (often sublingual area)

Relatively easy, evaluation of perfusion heterogeneity

Tissue PCO2

Adequacy of microvascular perfusion to metabolic needs

Numeric targets, almost continuous measurement


Microvascular reactivity

Easy, achievable in most patients

Affected by microvascular shunt Requires collaboration or sedation, not feasible in noninvasive ventilation, training of operator Impact of Haldane effect, not (yet?) widely available, costs (?), best organ/place to monitor? Microvascular reactivity may fail to reflect actual microvascular perfusion

CRT ¼ capillary refill time, SvO2/ScvO2 mixed-venous and central venous O2 saturation. VOT ¼ vascular occlusion test assisted either with laser Doppler or NIRS and evaluating microvascular reactivity.


D. De Backer, A. Durand / Best Practice & Research Clinical Anaesthesiology 28 (2014) 441e451

Mixed venous and central venous O2 saturation Mixed venous and central venous O2 saturations (SvO2 and ScvO2) reflect the balance between oxygen delivery and oxygen consumption. These are very useful to detect global decreases in tissue perfusion. Hence, both are important to monitor in the perioperative period [48] and in shock states [1]. However, when normal, ScvO2 and SvO2 cannot rule out any impairment in tissue oxygenation related to an impaired microcirculation. In tissues with heterogeneous perfusion, the over-perfused zones lead to an increase in O2 saturation in collecting venules, and hence in SvO2/ScvO2. In line with this concept, both low and elevated ScvO2 values are associated with a poor outcome in septic patients [49]. Accordingly, a low ScvO2/SvO2 suggests that tissue oxygenation may be compromised and that maneuvers aiming at improving oxygen transport should be considered, while a normal or elevated values/SvO2 may not be reassuring and should be interpreted in conjunction with other variables such as lactate and/or veno-arterial differences in PCO2. Direct measurements/estimates Doppler-related techniques Tissue blood flow can be determined with various laser Doppler techniques; these allow a dynamic evaluation of the flow present in a volume of tissue close to 0.5 mm3. Thus, the measured flow is the result of an aggregate flow in multiple small vessels, which does not enable the detection of microvascular heterogeneity. This technique can thus be useful when facing a homogeneous decrease in perfusion as in hemorrhagic shock (i.e., perioperative setting or trauma) but not in sepsis. Scanning laser Doppler and reflected-mode confocal laser scanning microscopy are recently developed laser techniques that allow the evaluation of perfusion and of its heterogeneity [50,51]. Due to the large size of the devices, these techniques can only be used to evaluate skin perfusion, even though these can be applied on most organs in the experimental setting. Microvascular reactivity, which reflects the capacity to recruit arterioles and capillaries after transient ischemia, may also be assessed with laser Doppler. In this test, arterial occlusion is caused by transient inflation of a cuff placed around the arm, the ascending slope observed just after the relief of the cuff reflects the quality of flow recovery [52]. Laser speckle imaging is a new modality that allows characterization of blood flow heterogeneity [53,54]. Red blood cells consist in individual moving scatters that alter the speckle pattern. These fluctuations provide information about the velocity of the scatters. The addition of scanning allows evaluation of blood flow heterogeneity. Using this technique, it has been feasible to evaluate heterogeneity in renal perfusion in septic animals [53] and skin perfusion in humans [54]. Videomicroscopy The application of a microscope on organ surfaces in humans requires specific techniques to illuminate the tissue without being polluted by light reflected by the surface. The first option is transillumination of thin organs, such as fingers in nailfold videocapillaroscopy [55], but its use is limited in critically ill patients due to peripheral vasoconstriction. Alternatively, organs can be made translucent by the reflection of light by deeper tissue layers. To discard the light reflected by tissue superficial layers, several options can be considered. With polarized light, the light reflected by deeper layers loses its polarized aspect and is discarded by filters. Other techniques optimize the geometry of the camera, separating emitting diodes and receiving optic paths. Orthogonal polarization spectral (OPS), sidestream dark-field (SDF), and incident dark-field (IDF) are three imaging techniques that can be applied at the bedside on critically ill patients with handheld devices. Red blood cells can be seen as dark/gray bodies as the used wavelength is absorbed by the hemoglobin. In septic animals, these have been used to evaluate the microcirculation of several organs as the tongue [10,56,57], gut [10,56,58], and brain [13,14]. In septic humans, these techniques have been mostly used to study the sublingual area [5,6,17,21,39,40]. Secretions and movement artifacts may impair image quality, and the investigation of the sublingual area is only feasible in sedated or cooperative patients. Vascular (all vessels and capillary) density and heterogeneity of perfusion (the proportion of perfused vessels, mean flow index, and heterogeneity index) should be measured [59]. The microvascular flow index (MFI) is a composite and nonlinear score

D. De Backer, A. Durand / Best Practice & Research Clinical Anaesthesiology 28 (2014) 441e451


that combines gross evaluation of perfusion and its heterogeneity. These are easily measured by experienced investigators using a semi-quantitative analysis, with excellent reliability [5,40,60]. The measurement of blood flow cannot be obtained with the scores in use. Computer-assisted microcirculation assessment allows the measurement of vessel density and blood flow in microvessels but still requires major human intervention. Measuring blood flow in selected microvessels is irrelevant, given the huge heterogeneity of blood flow in both normal and pathologic conditions. Measuring blood flow in all visible vessels is probably more relevant, but this is not yet feasible in clinical practice. Indirect measurements/estimates Near-infrared spectroscopy Evaluation of tissue oxygenation is an indirect evaluation of tissue perfusion, inferring the balance between O2 transport (DO2) and O2 consumption (VO2) in a tissue. Flow, hemoglobin content, arterial PO2, and VO2 are all of influence. Near-infrared spectroscopy (NIRS) uses near-infrared light to measure tissue oxy- and deoxyhemoglobin, and with some algorithms, myoglobin, and cytochrome aa3. Tissue O2 saturation (StO2) mostly represents the saturation of all vessels in the sampling volume, while total tissue hemoglobin (HbT) and the tissue hemoglobin index (THI) indicate the amount of blood present in the tested region. There are several limitations with NIRS. The first is that StO2 values are mostly influenced by local venous hemoglobin O2 saturation (~70% of StO2 measurement) and this proportion varies depending on the underlying condition. In hemorrhage, the venous component decreases markedly while arterial one is preserved. In trauma patients, StO2 is not very sensitive and was altered only in the most severe cases [61]. Another important limitation is that perfusion heterogeneity cannot be detected. The interpretation of StO2 is thus also difficult in sepsis as it does not reflect capillary O2 saturation nor does it correlate with ScvO2 saturation [62,63]. StO2 was demonstrated to be lower in septic patients compared to healthy volunteers, but with a major overlap between the groups [64]. Of interest, the NIRS technique allows the evaluation of the microvascular reactivity: a brief episode of forearm ischemia induced by transient inflation of a cuff (vaso-occlusive test, VOT) determines changes in StO2. Several indices can be measured during this test, but the most valuable is the ascending slope that reflects microvascular reactivity. The ascending slope is affected in critically ill patients, and especially in sepsis, and may be used to assess the response to therapy [8,64]. This provides quantitative information within few minutes and can be repeated and reproduced. However, the vasoreactivity test does not evaluate regional nor microvascular perfusion, but rather microvascular reserve. The thenar eminence is usually used for NIRS measurements as both adipose tissue thickness and edema can modify the collected data. NIRS-dynamic derived measurements (VOT test) demonstrated profound alterations in microvascular reactivity in septic patients compared to controls [8,64]. Similar alterations were reported in patients with severe heart failure [65]. The severity of these alterations in microvascular reactivity is associated with organ function [66,67], length of stay [67], and mortality [8,64]. These alterations can already be demonstrated in the Emergency Department and are more severe in non-survivors [8]. VOT appears helpful for monitoring and patients' management [65,67]. Tissue PCO2 and veno-arterial CO2 gradient Tissue PCO2 (PtCO2)-derived measurements can also be used to evaluate the microcirculation. The three major determinants of PtCO2 are PaCO2, VCO2, and the tissue blood flow. In normal conditions, an increased tissue metabolism (thus VCO2) is coupled with an increased tissue perfusion, largely reducing any PtCO2 increase (“washout” phenomenon). Therefore, if PaCO2 is constant, an increase in PtCO2 reflects an inadequate relationship between metabolism and tissue perfusion. PtCO2 thus represents a good estimate of tissue perfusion. To overcome the influence of PaCO2 on PtCO2, it is convenient to use the PCO2 gap (tissue-arterial CO2 gradient, normal

Monitoring the microcirculation in critically ill patients.

Alterations in microvascular perfusion have been identified in critically ill patients, especially in sepsis but also in cardiogenic shock, after card...
499KB Sizes 6 Downloads 13 Views