Towards integrative physiological monitoring of the critically ill: from cardiovascular to microcirculatory and cellular function monitoring at the bedside

Current hemodynamic monitoring of critically ill patients is mainly focused on monitoring of pressure-derived hemodynamic variables related to systemic circulation. Increasingly, oxygen transport pathways and indicators of the presence of tissue dysoxia are now being considered. In addition to the microcirculatory parameters related to oxygen transport to the tissues, it is becoming increasingly clear that it is also important to gather information regarding the functional activity of cellular and even subcellular structures to gain an integrative evaluation of the severity of disease and the response to therapy. Crucial to these developments is the need to provide continuous measurements of the physiological and pathophysiological state of the patient, in contrast to the intermittent sampling of biomarkers. As technological research and clinical investigations into the monitoring of critically ill patients have progressed, an increasing amount of information is being made available to the clinician at the bedside. This complexity of information requires integration of the variables being monitored, which requires mathematical models based on physiology to reduce the complexity of the information and provide the clinician with a road map to guide therapy and assess the course of recovery. In this paper, we review the state of the art of these developments and speculate on the future, in which we predict a physiological monitoring environment that is able to integrate systemic hemodynamic and oxygen-derived variables with variables that assess the peripheral circulation and microcirculation, extending this real-time monitoring to the functional activity of cells and their constituents. Such a monitoring environment will ideally relate these variables to the functional state of various organ systems because organ function represents the true endpoint for therapeutic support of the critically ill patient.

Shoemaker utilized the hemodynamic data obtained from the pulmonary artery catheter in high-risk adult surgical patients before, during and after surgical procedures. From these obser va tional data, he established a protocol using super normal values for cardiac output, oxygen delivery and oxygen consumption as the therapeutic goals. Indeed, this approach seemed to be favorable in surgical patients because it resulted in improved outcomes [3]. Donati and colleagues demonstrated that this approach was also successful in reducing morbidity and the length of hospital stay in high-risk surgical patients [4]. However, the eff ectiveness of this strategy in other critically ill patients remains controversial. Gattinoni and colleagues, for example, found no diff erence between patients who were treated with a protocol that targeted normal values for cardiac output, oxygen delivery and oxygen consumption, super normal values or mixed venous saturation >70% [5]. Hayes and colleagues found an increased mortality in patients who were treated using the supranormal values protocol [6]. Th e condition of normal or reduced oxygen extraction seemed to be a key hemodynamic component that contraindicated targeting supranormal values of oxygen delivery. Vincent suggested that the application of a dobutamine challenge to identify the eff ect of increasing systemic oxygen delivery was an eff ective strategy to achieve hemodynamic optimization for the critically ill patient [7].

Adding the peripheral circulation to the equation
Simply targeting the parameters related to the systemic circulation was ineff ective in resuscitating the various organ systems because the microvasculature was unable to eff ectively regulate the fl ow of oxygen-carrying blood to match regional needs. Th is inability is presumably caused by the pathogenic action of the infl ammatory mediators, reactive oxygen species and hypoxemia on vascular regulatory mechanisms, such as autoregulation. Th is dysfunction is augmented by certain therapies such as intravascular fl uids and vasoactive mediators that override the endogenous physiological mechanisms regulating homeostasis. Th e consequent mismatch has its impact not only between diff erent organ systems but also at the level of the microcirculation. Th is defect manifests itself as a reduced oxygen extraction defi cit that is characterized by shunting within vulnerable, weak microcirculatory units and organ beds. Reduced extrac tion results in elevated venous oxygen levels in the presence of signs of regional dysoxia, such as elevated levels of lactate and elevated tissue carbon dioxide (CO 2 ) [8].
Weil, in advance of experimental evidence, fi rst understood the origin of the sequence of events relating to oxygen transport dysfunction during circulatory failure [9,10]. He classifi ed four states of shock: hypovolemic, obstructive, cardiogenic and distributive. All of these states indicate that the ultimate target of shock is the cellular starvation of oxygen availability. Th e fi rst three states of shock, however, are associated with the reduction in cardiac output that is the primary cause of the ensuing tissue dysoxia. Distributive shock can occur in the presence of a normal or even elevated cardiac output and describes a defect in the vascular traffi cking of cardiac output between and within the various organ beds, resulting in local tissue dysoxia in the presence of otherwise normal systemic hemodynamics. When such a distributive defect occurs, increased cardiac output is ineff ective in resolving the regional tissue dysoxia. In their editorial 'Expanding from the Macro to the Microcirculation' , Weil and Tang identifi ed the need to monitor the microcirculation to monitor and treat distribu tive shock [11]. Th is realization, in combination with the idea that microcirculatory dysfunction leads to organ dysfunction, formed the basis for the appreciation of the microcirculation as a central focus in the pathogenesis of multiorgan dysfunction [12].
Th e need to monitor hemodynamic and oxygenderived variables of the peripheral circulation was demon strated by the introduction of CO 2 gastric tonometry by Fiddian-Green and Baker to identify splanchnic dysoxia during states of shock [13]. Th e signifi cance of this monitoring for the critically ill was revealed by the landmark study of Guteriez and coworkers, who found that septic shock patients whose gastric CO 2 did not normalize following resuscitation had a higher chance of dying than did those whose gastric CO 2 normalized [14]. Th e physiological mechanisms underlying tissue CO 2 production was controversial at that time, however, with one school of thought indicating that its origin was due to mitochondrial dysfunction and another school support ing the idea that abnormal CO 2 refl ected an abnormal perfusion of the tissues [15]. However, a number of experimental studies performed by Dubin and coworkers [16] as well as clinical investigations by Creteur and colleagues [17] have now fi rmly established that elevated tissue CO 2 refl ects a perfusion defi cit in the microcirculation. Th e importance of monitoring the peripheral circulation in the context of distributive shock was further expanded upon by the work of Lima and Bakker, who investigated the clinical signifi cance of assessing peripheral perfusion by physical examination. In so doing, Lima and colleagues identifi ed abnormalities in peripheral perfusion as being associated with high Sequential Organ Failure Assessment scores [18].

A closer look at the microcirculation
Clinical monitoring of the microcirculation has previously been limited to indirect measures such as lactate, tissue CO 2 and subjective assessment of peripheral per fusion. Hand-held intravital microscopes off er a diff er ent approach [19,20], which incorporates specialized optics such as crossed polarized green light and/or dark-fi eld illumination to fi lter out the surface refl ections as developed much earlier by Slaaf and colleagues and Sherman and colleagues [21,22]. Th is technology allows for observation of the fl owing red blood cells in the microcirculation of the mucosal surfaces of organ beds [19,23]. Th ese hand-held intravital microscopes were subse quently used to directly observe the micro circulation on organ surfaces at the bedside in various clinical scenarios [24][25][26][27][28][29][30][31][32][33]. Sublingual microcirculatory observations identifi ed micro circulatory obstructions to be characteristic of septic patients who are resistant to therapy despite corrected systemic hemodynamics [25,32,33]. De Backer and coworkers fi rst demonstrated a correlation between the severity of microcirculatory alterations and morbidity and outcome in septic patients, whereas no such relation ship existed for conventional systemic hemodynamic variables [26]. We recently further demon strated this phenomenon in septic pediatric patients [34]. Th ese fi ndings were reproduced using an early goal-directed therapy treatment by Tryziack and coworkers, who found that an early eff ective recruitment of the microcirculation predicted Sequential Organ Failure Assessment improvement 24 hours following early goal-directed therapy [25].
Based on the idea that active recruitment of the microcirculation is needed for resuscitation, vasodilatory therapy (for example, nitroglycerin) was shown to be especially eff ective in recruiting obstructed sublingual microcirculation in pressure-resuscitated septic patients [24]. Similar improvement was not found in fl uidoptimized septic patients [35]. In septic shock patients, levosimendan was demonstrated superior to dobutamine for recruiting microcirculation [36], while the addition to norepinephrine of continuously infused low-dose of terlipressin or vasopressin did not aff ect sublingual micro circulatory blood fl ow [37].
In cardiac surgery patients, blood transfusions are eff ective in improving microcirculatory oxygen availability by recruiting previously unfi lled microcirculatory capillaries, thereby reducing the diff usion distances between capillaries and tissue cells; this result emphasizes the importance of viscosity in recruiting the microcirculation during resuscitation [38]. Th e importance of viscosity was further demonstrated in a study of septic patients by Dubin and colleagues, who showed that highly viscous starch solutions can recruit the microcirculation more eff ectively than less viscous crystalloids [39]. Th ese studies highlight the unique ability of microcirculatory monitoring to measure not only fl ow (convection) but also the diff usive capacity of the circulation to transport oxygen by measuring the func tional capillary density [31,39,40].
In studies of septic patients, fl uid responsiveness has been evaluated at the level of the microcirculation. Th e type and timing of fl uid administration have been found to be an important aspect of fl uid effi cacy in recruiting microcirculation [39,41]. Vasopressor therapy, although eff ective in increasing blood pressure, can have limited or even deleterious eff ects on improving perfusion of the microcirculation [30,42]. One consistent fi nding from various investigators has been that microcirculatory alterations often manifest themselves at the capillary level by normalized or even elevated fl ow in the larger venules [24,26,33]. Th ese observations describe the nature of the distributive defect that occurs during shock (especially during the resuscitation phase, as obstructions in the capillary vessels aff ect the persistence of fl ow in the larger microvessels) and, furthermore, directly illustrate the nature of the functional shunting that is associated with sepsis and other forms of distributive shock [8]. In particular, the heterogeneity of capillary function has been found by many to be a key characteristic feature of this type of distributive shock [33]. Th is observation led Tryziack and co-workers to analyze microcirculation images to develop a heterogeneity index to quantify this type of microcirculatory alteration [25].
An additional level of heterogeneity can be attributed to the physiological diversity of the patients themselves. In particular, diff erences in age infl uence the response of the patients, with each age group having its own characteristic phenotype and response to critical illness. In this respect, critically ill pediatric and neonatal patients form a special group because they present a completely separate level of (patho)physiological diversity relative to adult patients [43]. For example, as an infant grows during the fi rst years of life, systolic and diastolic pressures are low and heart rates are high. Th e cardiac output and stroke volume continue to rise until the age of 5 years. Changing cardiovascular physiology is also refl ected in the development of the microcirculation, which exists during the initial days and months following birth as a rich network of microcirculatory capillaries that diminishes in density as the infant grows [44].
Th e response to critical illness is also largely divergent between pediatric and adult patients. A diminished systemic vascular resistance is a hallmark of adult sepsis but is not observed in the pediatric patient. Furthermore, septic shock in pediatric patients, in contrast to adult patients, is often characterized by a hypodynamic response with low cardiac output and high systemic vascular resistance, although a rapid switch can be made. Th e septic pediatric patient also has a diminished contractile reserve and a poor response to volume loading and inotropic support [43]. Hemodynamic monitoring in these very small patients is indeed a challenging task because the possibilities for invasive hemodynamic monitor ing are limited. For instance, Swan-Ganz monitor ing has never become common practice in the pediatric age group. Hemodynamic monitoring using hand-held intravital microscopes could off er advantages in these patients; besides targeting an important physiological compartment, this method off ers the additional advantage of being largely non-invasive.
Th e potential application of monitoring the microcirculation in pediatric patients using hand-held intravital microscopic techniques has been exemplifi ed in the work of Top and Tibboel. Using orthogonal polarization spectral imaging, Top and colleagues demonstrated in septic children that persistent microcirculatory alterations were the single most sensitive and specifi c indicator of outcome [34]. In a recent study, Paize and coworkers further supported the importance of monitoring microcirculatory altera tions in pediatric patients by observing that certain micro circulatory alterations in patients with severe meningo coccal disease are associated with clinical recovery [45]. Others have shown that some therapies, such as hypothermia and blood transfusions, positively impact the microcirculation of critically ill pediatric and neonatal patients [46,47].
Th e cardiovascular response of neonatal patients is a largely unexplored area and presents a further level of complexity, not only owing to their small size but also owing to their complex response to hypoxia [48]. Th is issue is truly a physiological challenge. Whereas hypoxemia is considered a pathological condition in older patients, hypoxia may be viewed as a physiological condition for the neonate, to which the neonate is continuously adapting. Only when these adaptive mechanisms fail does the neonate present as critically ill. Th ese adaptive mechanisms require support that is essential for promoting development [49]. Monitoring the success of the microcirculation at providing blood fl ow in combination with an assessment of tissue oxygenation is anticipated to form an important platform to realize this support.

Cell function monitoring
Th e microcirculation is an integrative physiological compartment in which red blood cells, leucocytes, blood constituents, endothelial cells, smooth muscle cells, paren chymal cells and the intracellular components of these cellular systems integratively and symbiotically function together to ensure optimal oxygen and nutrient transport for the utilization of the parenchymal cells. Indeed, adequate function in terms of perfusion and oxygen transport can be regarded as an indication of success for all of these cellular systems. Microcirculatory dysfunction of this system caused by pathogenic factors, such as infl ammation, oxidative stress and hypoxemia, however, can lead to organ dysfunction [12]. Fully understanding the nature of the insult and the indication for appropriate therapy requires insight into the function of the individual subcellular building blocks of the microcirculation. Th e future of monitoring will need to integrate the functional state of the various cellular constituents into micro circulatory monitoring.
Th e ability of red blood cells to carry hemoglobinbound oxygen to the microcirculation is, of course, one of the main functions of the cardiovascular system. Th e oxygenation state of hemoglobin in red blood cells can be measured quite eff ectively at the bedside using spectrophotometry, and we used this method to demonstrate the effi cacy of blood cell transfusion to improve oxygen availability in the microcirculation in adult anemic hematological patients [50]. Leukocytes form an important source of pathogenic activation, resulting in tissue damage that contributes to organ dysfunction. Th e ability to monitor leukocyte activation at the bedside using direct obser vation of their rolling and sticking to the endothelium could therefore provide an important indication of the state of infl ammation and possibly the response to therapy. Hand-held intravital microscopy of the sublingual bed was fi rst used for this purpose by Baur and coworkers in patients following the release of the clamp after cardiac surgery [51].
Th e endothelial cell forms the central regulatory player in the orchestration of the physiology of microcirculation. Th e cell plays an important signaling role in the regulation of vessel tone in addition to infl ammation and hemostasis. Assessment of its function can be accomplished by administering compounds that target endothelial cell function and observing the microcirculatory response. De Backer and coworkers used this approach by topically administering acetylcholine sublingually in septic patients who demonstrated enhanced microcircula tory perfusion [52].
A critical subcellular component that has come to prominence recently due to its relevance to critical illness is the endothelial glycocalyx [53]. Th is gel-like layer lining the endothelial cells forms the barrier between the intravascular lumen and the endothelial cells. Shedding and disruption of the glycocalyx has been associated with many states of endothelial dysfunction, including the loss of auto regu lation and the development of tissue edema and organ dysfunction. Vink and coworkers developed a method to measure the integrity of the glycocalyx by analyzing images obtained from sublingual intravital microscopy [54]. Th ey further developed a software platform to assess the functional state of the glycocalyx directly at the bedside [55].
Th e routine clinical application of such measurements using the current orthogonal polarization spectral/ sidestream dark-fi eld hand-held intravital microscopes and analog video cameras [19,23] has been criticized [56,57]. Th is criticism is based on the fact that these devices have poor reproducibility and image quality [58,59] and require time-consuming off -line analysis of the acquired images. Th ese devices also suff er from pressure artifacts imposed by the weight of the devices [56,57] and from an inability to implement automatic analysis software to process the generated images [60]. In addition, higher resolution optics and image sensors are required to allow for software analysis and the identifi cation of the subcellular structures associated with microcirculatory function. For these devices to enter the clinical arena, technological advances are therefore mandatory [56,57].
A new hand-held intravital microscope has been developed recently that is based on incident dark-fi eld imaging [22], containing a computer-controlled highresolution imaging sensor [61]. Such technological advances might possibly address the earlier critiques of the conventional devices but will need to be validated with regard to these critiques before such hand-held vital microscopes can truly enter the clinical arena.

Towards an integrated physiological monitoring system
Th e above summary has highlighted the need to extend monitoring of the physiological determinants of organ function from the macro to the micro and down to the cellular level. However, a crucial component of this monitoring is the need to include functional indicators of organ function because it is the successful restoration of organ function that determines the success of intensive care. Th ese indicators of organ function need to be continuous, specifi c and quantitative. Th ese initiatives are important because they describe a road map for the new developments that are needed to provide complete physiological monitoring of critically ill patients. Th e information from new sensors and physiological variables, as well as measures of organ function, will require a much higher level of integration than is currently available, and mathematical models of physiology and pathophysiology are expected to play an important role in this integration. From this perspective, these innovations represent a challenge for industry.
By integrating information on all of the characteristics of the patient -including disease, co-morbidities and age -into the evolution of this integrated physiological monitoring system, we anticipate the development of an environment in which the complete continuum of human development, as well as diseases and their response to therapy, can be monitored. Intensive care medicine off ers a unique environment for this development, which ultimately may be relevant to other areas of medicine.