Complications after major surgery are a leading cause of morbidity and mortality. The etiology of postoperative complications is complex, but poor cardiorespiratory reserve appears to be a key factor. There is increasing interest in the use of central and mixed venous oxygen saturation to guide therapeutic interventions during the perioperative period. However, a detailed understanding of the physiologic principles of venous oximetry is essential for safe and effective use in clinical practice. Venous oxygen saturation reflects the balance between global oxygen delivery and oxygen consumption, which may be affected by a wide range of factors during the perioperative period. The purpose of this article is to describe the physiology and measurement of mixed and central venous oxygen saturation and to explore the findings of clinical investigations of their use in perioperative care.
IT is estimated that 234 million major surgical procedures are performed worldwide each year.1Complications after major surgery are a leading cause of morbidity and mortality. High-risk surgical patients account for more than 80% of deaths but less than 15% of in-patient procedures.2,3Data from across the developed world confirms that poor outcomes after high-risk surgery are a global problem.4–6Even for those patients who survive to leave hospital, postoperative complications remain an important determinant of long-term survival.6It is therefore essential that we seek to improve outcomes for patients undergoing major surgery.
The etiology of postoperative complications is complex, but poor cardiorespiratory reserve appears a key factor. A number of reports indicate that poor outcomes after major surgery are strongly associated with derangements in tissue oxygen delivery that may in turn relate to impaired microvascular flow.7–10The use of fluid and inotropic therapy to enhance tissue oxygen delivery may reduce the incidence of postoperative complications.11–14There is an increasing body of literature describing changes in central (Scvo2) and mixed venous oxygen saturation (Svo2) during the perioperative period, which, along with a recent study in patients with severe sepsis,15has led to interest in the use of venous saturation as a therapeutic goal for surgical patients. However, the complexities of the physiology of venous oxygen saturation are poorly recognized. A detailed understanding of these principles is essential for the safe and effective application in clinical practice. The aim of this article is to describe the physiology and measurement of Svo2and Scvo2and to describe the findings of the clinical investigations of the use of these variables in perioperative care.
Materials and Methods
Searches of the MEDLINE and Cochrane CENTRAL databases from January 1968 to December 2008 were performed by both authors using the following search terms: (venous saturation OR venous oximetry OR Svo2OR Scvo2) AND (surgery OR surgical OR \ *operative OR operation). Only articles published in English were included, but no restrictions were placed on source. A further online search was then carried out using the Google Scholar search engine by using the following key words: venous saturation, venous oximetry, surgery, Scvo2, Svo2. The resulting abstracts were screened to identify relevant investigations in adult patients undergoing major surgery. Studies were excluded if they had not been published in a peer-reviewed journal. Bibliographies of relevant articles were also screened. Manuscripts were screened initially by title and then by abstract before obtaining the full text of relevant articles.
Physiology of Venous Oxygen Saturation
The terms central (Scvo2) and mixed venous oxygen saturation (Svo2) refer to the hemoglobin saturation of blood in the superior vena cava and proximal pulmonary artery, respectively.16Rearrangement of the Fick equation illustrates that venous oxygen content is determined by arterial oxygen content, oxygen consumption and cardiac output.17The quantity of dissolved oxygen is small under standard conditions; therefore, the more conveniently measured variable of hemoglobin saturation is preferred. This is summarized in the equation below, where CO refers to cardiac output, Cao2refers to arterial oxygen content, Cvo2refers to venous oxygen content, and Vo2refers to oxygen consumption.
Where oxygen supply is insufficient to meet metabolic requirements, increased tissue oxygen extraction results in a decrease in the oxygen content of effluent venous blood. Venous oxygen saturation therefore reflects the balance between global oxygen delivery (Do2) and global oxygen consumption (Vo2).18Vo2and Do2both fluctuate significantly during the perioperative period, and it is of particular importance to recognize that changes in venous saturation may reflect a variety of physiologic and pathologic changes (fig. 1). The safe use of venous saturation as a therapeutic goal depends on the prompt recognition of the cause of any derangement. Regional variations in Do2and Vo2are also commonplace and clinically relevant differences in the oxygen content of venous blood are to be expected in different parts of the circulation.19–22In common with other global physiologic variables, the apparent simplicity of a single variable is often associated with a lack of sensitivity to detect regional abnormalities in an apparently stable patient. There is little published data describing the normal value of venous saturation in health. Although commonly quoted as 70%, the available data suggest this may vary from 70% to 80% in healthy individuals.23,24Values of Svo2and Scvo2may often be as low as 65% in hospital in-patients before elective surgery.25
Oxygen Delivery as a Determinant of Venous Oxygen Saturation
Global oxygen delivery is determined by cardiac output and the oxygen content of arterial blood as shown in the equation below,26where Do2refers to oxygen delivery and CO to cardiac output where the Bunsen solubility coefficient for O2at 37°C is 0.02.27
Adequate tissue oxygen delivery therefore depends on the adequacy of both respiratory and cardiovascular function. If oxygen consumption, hemoglobin concentration, and arterial saturation remain constant, changes in Svo2are therefore directly proportional to those in cardiac output; this relationship has been demonstrated in several studies in man.24,28In a study of healthy volunteers, orthostatic hypotension resulted in a decrease in cardiac output from 4.3 to 2.7 l min−1at the onset of presyncopal symptoms.24Over the same time period, Scvo2decreased from 75% at baseline to 60%.24A clinical series of patients undergoing one-lung ventilation demonstrated that cardiac output increased in response to sudden decreases in arterial saturation; as a consequence, Svo2remained unchanged.28Several reports describe reduced venous saturation in patients with a reduced cardiac output due to myocardial infarction and/or heart failure.29–34Changes in Scvo2and Svo2in these circumstances reflect both the severity of hemodynamic disturbances and response to treatment.29–32
The affinity of hemoglobin for oxygen is affected by the partial pressure of oxygen (fig. 2). It may be anticipated from the oxyhemoglobin dissociation curve that, at higher partial pressures of oxygen, increases in Po2will result in only small increases in hemoglobin saturation. At lower partial pressures, such as those typical of venous blood, the same incremental rise in Po2will result in a greater increase of hemoglobin saturation due to the greater oxygen affinity of deoxyhemoglobin.35Consequently, the change in venous saturation in response to a step change in fractional inspired oxygen concentration may differ considerably from simultaneous changes in arterial hemoglobin saturation. Clinical and laboratory investigations have shown that an increase in fractional inspired oxygen concentration results in a greater increase in oxygen saturation of venous than arterial blood.36,37The administration of supplemental oxygen may therefore be sufficient to rectify significant abnormalities in venous saturation even though these abnormalities may not specifically result from alveolar hypoxia. In situations where Scvo2or Svo2values are being used as hemodynamic endpoints for the administration of intravenous fluid or inotropic therapies, an increase in venous saturation resulting from an increase in fractional inspired oxygen concentration may be misinterpreted as an indication of adequate hemodynamic resuscitation. The potential for such simple interventions to mask the effects of shock emphasizes the importance of a detailed understanding of venous oximetry.
Oxygen Consumption as a Determinant of Venous Oxygen Saturation
Few studies have explored the relationship between Vo2and venous saturation during the perioperative period. This may reflect poor recognition of the importance of Vo2as a determinant of venous saturation. Considerable changes in oxygen consumption may occur during the perioperative period. Increases in Vo2resulting from pain, anxiety, or shivering may all result in a decrease in venous saturation,38–42whereas the corresponding treatments may rectify such derangements.43Experimental data suggests that the extent of such derangements may correlate with the magnitude of oxidative stress.44General anesthesia results in a decrease in Vo2through reductions in general motor activity, work of breathing, neuronal activity, and body temperature. These changes are the result of anesthesia itself as well as neuromuscular blockade and invasive ventilation.45–49Volatile anesthetic agents decrease the basal metabolic rate, with reductions in sympathetic tone and cardiac output being more pronounced at higher doses.50Intravenous hypnotics such as benzodiazepines appear to exert similar effects on metabolic demand by blunting the sympathetic neurohumoral response,51,52and intravenous anesthetic agents such as propofol similarly reduce metabolic demand, with the probable exception of ketamine, which usually increases myocardial inotropy by increasing general sympathetic activity.53,54Sympatheticolytic agents such as clonidine reduce perioperative Vo2.55,56Reductions in neuronal oxygen consumption occur with the administration of volatile anesthetic agents, barbiturates, benzodiazepines, and propofol.57–61Opiates may similarly reduce perioperative Vo2.62–65Neuraxial blockade has both sympatheticolytic and analgesic effects,66,67but we are unaware of reports specifically describing effects on Vo2.
Relationship between Svo2And Scvo2
While the determinants of Scvo2and Svo2are very similar, the relationship between the two variables is complex and they cannot be used interchangeably.68–73Regional variations in the balance between Do2and Vo2result in differences in the hemoglobin saturation of blood in the superior and inferior vena cavae.74Streaming of caval blood continues within the right atrium and ventricle and complete mixing only occurs during ventricular contraction. The drainage of myocardial venous blood directly into the right atrium via the coronary sinus and cardiac chambers via the Thebesian veins results in further discrepancies.16Consequently, Svo2reflects the balance between oxygen supply and demand averaged across the entire body but Scvo2is affected disproportionately by changes in the upper body.74In healthy individuals, Scvo2is usually 2–5% less than Svo2,16largely because of the high oxygen content of effluent venous blood from the kidneys.22This relationship changes during periods of hemodynamic instability because blood is redistributed to the upper body at the expense of the splanchnic and renal circulations.75In shock states, therefore, the observed relationship between Scvo2and Svo2may reverse, and the absolute value of Scvo2may exceed that of Svo2by up to 20%.73This lack of numerical equivalence has been demonstrated in various groups of critically ill patients, including those with cardiogenic, septic and hemorrhagic shock.31,68–70,76–78This has also been demonstrated in patients undergoing general anesthesia for cardiac71,72,79and noncardiac surgery.69,80Although trends in Scvo2may closely reflect those of Svo2, absolute values differ and the variables cannot be used interchangeably.68–72This observation is sometimes cited in support of continued use of the pulmonary artery catheter. However, there is no evidence to suggest one variable is of greater clinical value than the other. As the use of the pulmonary artery catheter declines, measurement of Scvo2is usually more convenient than Svo2, although Scvo2measurements cannot be cannot be used to calculate Vo2or shunt fraction.72
Measurement of Venous Oxygen Saturation
Cardiac catheterization was first performed in 1929 by Werner Forssmann, a major advance that allowed the measurement of Svo2and hence application of Fick’s principle to measure cardiac output.81,82However, it was not until 1970 that the introduction of the balloon-tipped pulmonary artery catheter facilitated the routine clinical measurement of Svo2.81Reports of the clinical utility of Scvo2predate those of Svo2by several years.29,33Measurement of venous saturation may be performed either intermittently by blood sampling and cooximetry or continuously through the use of a spectrophotometric catheter.
Intermittent Blood Sampling and Cooximetry
Cooximetry involves the measurement of hemoglobin saturation by spectrophotometry by using widely available blood gas analysis technology. The differences in light absorption spectra between oxygenated and deoxygenated hemoglobin allow calculation of the hemoglobin saturation of blood. This also allows the identification of other forms of hemoglobin such as methemoglobin and carboxyhemoglobin. Cooximetry is a reliable and well-established technique. However, in clinical practice it may be inconvenient to make frequent measurements by using this approach. Specific errors result from sample contamination, delayed measurement, and sampling from the incorrect site.83,84As with any form of venous oximetry, interpretation errors may arise due to intracardiac shunts, tricuspid regurgitation, and catheter misplacement.84When taking blood samples, syringe aspiration should be gentle enough to avoid high negative pressure that may increase the aspiration of pulmonary capillary blood and hence produce falsely high readings for oxygen saturation.
Continuous Measurement Using an Indwelling Fiberoptic Catheter
The introduction of optical fiber technology has allowed the continuous measurement of venous saturation by spectrophotometry using indwelling pulmonary artery or central venous catheters. The major benefit of this approach is the provision of continuous data allowing the detection of sudden fluctuations in venous saturation, which are common during the perioperative period.85,86The principle disadvantages of this technology are the additional cost and signal drift, although the latter can be addressed by recalibration. Advances in the technology have addressed the problem of interference from other optically active compounds such carboxyhemoglobin and bilirubin.
Observational Studies of Perioperative Changes in Scvo2and Svo2
Abnormalities of venous saturation are common during and after major surgery and are associated with an increased incidence of postoperative complications.87–91Reductions in Scvo2and Svo2also have prognostic significance in heart failure, trauma, and sepsis.92–95These observations are no surprise, given the wide range of pathologic abnormalities that affect venous saturation in the perioperative period.87–89
Two studies have been performed in noncardiac surgical patients with complementary findings. In the first observational study of 117 patients, the lowest recorded value of Scvo2in the early postoperative period was independently associated with subsequent complications, the optimal cut-off for the lowest Scvo2value being 64.4%.87Interestingly, a considerable decrease in Scvo2was observed within the first hour after surgery, possibly as a consequence of increased Vo2after the cessation of general anesthesia (fig. 3). In a further multicenter observational study of 60 patients, the mean value of Scvo2was found to be reduced at various time points throughout the perioperative period in patients who developed complications.88The optimal cutoff value in this study for the mean Scvo2value was 73%. These investigations not only provide strong evidence to support the role of Scvo2as a therapeutic target, but they are also highly consistent in suggesting the most appropriate target value to be an Scvo2value of approximately 75%. However, these findings do not indicate how venous saturation should be used as a therapeutic goal. A range of factors influence Vo2, Do2, and therefore venous saturation during the perioperative period, not all of which are pathologic in nature. The most appropriate therapy to achieve a venous saturation endpoint may vary.
Alterations in Svo2have been described in patients undergoing cardio-thoracic surgery, although no reports of changes in Scvo2were identified.89–91Derangements in Svo2occur before any changes in mean arterial pressure or heart rate are observed,96and they appear to correlate well with changes in cardiac index.86Early work in patients undergoing both cardiac and pulmonary surgery demonstrated that sustained reductions in Svo2below 65% were associated with a higher incidence of complications, particularly arrhythmias.97Increases in oxygen extraction ratio, derived through measurement of Svo2, have also been associated with postoperative organ failure and prolonged intensive care stay.90,91,98During lung transplantation, changes in Svo2reflected adverse clinical events, although this series is too small to support any more detailed conclusions.99During cardiopulmonary bypass, Svo2may prove a more specific indicator of global oxygen delivery; pump flow (or cardiac output) and metabolic rate are generally constant in these circumstances.100,101
The effects of hypovolemia on venous saturation have been described in both animals and humans.73,102,103Fluctuations in Svo2and Scvo2closely mirror periods of hemorrhage and subsequent resuscitation in anesthetized dogs.73,102A case series of ten victims of mainly penetrating trauma described similar changes in Svo2.103Venous saturation may provide a useful indication of the severity of blood loss that is more reliable than conventional cardiovascular variables such as heart rate and arterial and central venous pressure.102,103A single small case series describes the use of normal levels of Svo2as therapeutic target in trauma patients in which the authors suggest a survival benefit.104However, the study has a number of limitations, and the data do not appear to support such conclusions.
Interventional Trials Utilizing Scvo2and Svo2as Therapeutic Targets in the Perioperative Period
Our literature search identified only one interventional trial using Scvo2as a therapeutic goal in perioperative care.105This was a multicenter trial of 135 patients undergoing major abdominal (including aortic) surgery. All patients received fluid challenges, dobutamine up to 15 μg · kg−1· min−1and blood transfusions to achieve predefined goals for arterial pressure, urine output, and central venous pressure.105These same therapies were administered in the intervention group to achieve the additional goal of an estimated oxygen extraction ratio of less than 27%, the value of which was calculated using intermittent measurements of Scvo2. Trial interventions were continued until an unspecified time on the first postoperative day. Dobutamine was administered more frequently and in greater doses to the Scvo2group (2.6 ± 4.0 μg · kg−1· min−1vs. 0.4 ± 2.2 μg · kg−1· min−1; P = 0.001). Volumes of intravenous fluid and transfused blood were similar in the two groups, although fluid challenges were commenced earlier stage in the Scvo2group. Fewer patients in the Scvo2group developed organ failure (8 of 68 patients [11.8%]vs. 20 of 67 patients [29.8%]; P < 0.05). The duration of hospital stay was also reduced in the Scvo2group (11.3 ± 3.8 days vs. 13.4 ± 6.1 days; P < 0.05), whereas mortality was low in both groups (2.9% vs. 3.0%; absolute values not reported). This was an important investigation with encouraging findings. However, there are some limitations that prevent full interpretation of the results. The report provides little information regarding the standardization of interventions that are frequent confounders in trials of this size. In particular, there is little or no description of those interventions likely to limit excessive Vo2. These include anesthesia, analgesia, temperature maintenance, postoperative sedation, ventilation, and other aspects of postoperative critical care. It is unclear why the investigators chose to use estimated oxygen extraction ratio as a hemodynamic goal rather than absolute values of Scvo2. Although this may reduce the effects of alveolar hypoxemia as a confounder, the use of Scvo2to calculate oxygen extraction ratio is considered unreliable.68,73,76,80,106In common with a number of similar trials, the small sample size limits the generalizability of the findings.11–13Although the multicenter design offsets this somewhat, much larger trials are clearly needed to resolve the question of effectiveness in routine clinical practice.
In an earlier study of patients undergoing peripheral vascular surgery, the use of Svo2as a therapeutic endpoint for inotropic therapy was not associated with any change in outcome.107Patients undergoing aortic reconstruction or limb salvage procedures were admitted to intensive care 12 hours preoperatively for pulmonary artery catheter placement. Initial values of Svo2were surprisingly low but responded significantly in the intervention group (59.1% to 68.8%). However, final Svo2values were similar in the two groups (70.0% vs. 70.1%) perhaps explaining the similar outcomes.
Polonen et al. randomized 196 patients undergoing elective cardiac surgery to a protocol involving the administration of intravenous fluid and inotropic therapy to attain a target Svo2of at least 70% in the first 8 h after surgery.108Dobutamine was administered in doses of up to 15 μg · kg−1· min−1where the target Svo2was not achieved with intravenous fluid alone. Control group patients were administered intravenous fluid and dobutamine to meet goals for pulmonary artery occlusion pressure, cardiac index, arterial pressure, and hematocrit. Svo2was similar in the two groups at baseline (control group 67 ± 6% vs. Svo2group 67 ± 6%), but there were greater improvements in Svo2in the Svo2group (control group 69 ± 5% vs. Svo2group 71 ± 4%; P < 0.001). Svo2-guided therapy was associated with a reduction in both hospital stay (7 [5–8] days vs. 6 [5–7] days; P < 0.05) and the number of patients developing complications (11 patients [5.6%]vs. 2 patients [1.0%]; P < 0.01). It is uncertain whether such a small mean difference in Svo2of 2% is a true reflection of these improved clinical outcomes. In common with other trials, the intervention protocol principally targeted Svo2by increasing Do2. In addition, the authors report measures in all patients that would have minimized excessive Vo2. These include postoperative sedation and ventilation that was discontinued only when the patient was normothermic and hemodynamically stable. Hemodynamic therapy to attain a target value for Svo2is more appropriate in this context as confounding causes of decreased venous saturation are minimized. This treatment approach is possible after cardiac surgery where postoperative intensive care admission is a standard of care; this is not always the case for high-risk noncardiac surgery.2,3
Scvo2and Svo2reflect important pathophysiological changes in oxygen delivery and consumption that occur during the perioperative period. The most appropriate clinical interventions to rectify abnormalities of venous saturation may therefore vary widely. Supplemental oxygen, respiratory support, blood products, intravenous fluid, inotropic therapy, anesthesia, analgesia, sedation, and rewarming are all commonly used perioperative interventions that affect venous oxygen saturation. Small clinical trials suggest that the use of venous saturation as a therapeutic goal for hemodynamic therapy may reduce postoperative complication rates. However, these studies are not large enough to demonstrate a mortality benefit and are poorly generalizable. Further research is required to establish the most appropriate treatment algorithms for the use of Scvo2and Svo2in perioperative care. Large, prospective, randomized control trials should then be undertaken to confirm the effects of such an approach on clinical outcomes.