Secondary Logo

Journal Logo

Basic Science Aspects

Peripheral Muscle Near-Infrared Spectroscopy Variables are Altered Early in Septic Shock

Orbegozo, Diego; Su, Fuhong; Xie, Keliang; Rahmania, Lokmane; Taccone, Fabio Silvio; De Backer, Daniel; Vincent, Jean-Louis; Creteur, Jacques

Author Information
doi: 10.1097/SHK.0000000000000991



Despite advances in understanding and treatment, mortality in sepsis remains unacceptably high (1). The aim of protocolized strategies for the treatment of sepsis is to rapidly restore systemic hemodynamics and optimize indexes of global perfusion, such as central venous oxygen saturation. However, recent large multicenter trials have shown disappointing results using this approach (2). Indeed, the distribution of cardiac output is altered in septic states in an attempt to preserve perfusion to vital organs, like the brain and the heart, at the expense of the gut and the muscle (3–5), thus targeting systemic hemodynamics may not be optimal. In addition, even when systemic hemodynamics are restored, areas of poor perfusion and oxygen supply associated with a significant degree of shunting of the microcirculation are still present (6, 7). These alterations can lead to the development of organ dysfunction and increase the probability of poor outcome in such patients (8, 9).

Near-infrared spectroscopy (NIRS) is an attractive method for evaluating peripheral tissue perfusion and microvascular reactivity. This noninvasive technique estimates the tissue oxygen saturation (StO2) as a measurement of the microvascular oxygen content (10); StO2 can be assessed continuously at the bedside and its values are mainly affected by changes in the venular oxygen saturation (11, 12). However, normal values vary widely making it difficult to determine an optimal target to indicate adequate tissue perfusion during resuscitation of patients with sepsis (13, 14). As such, a transient vascular occlusion test (VOT), coupled with continuous StO2 measurements, enables better evaluation of microvascular function and peripheral muscular oxygen consumption. Indeed, the speed of decay of the StO2 signal is a reflection of local oxygen consumption, and the speed of reperfusion is a reflection of postocclusive reactive hyperemia (vascular reactivity) (15). These variables have been correlated with the severity of septic shock on ICU admission and remained independent predictors of poor outcome even after correction for other hemodynamic covariates (10, 16).

To study the time-course of these variables during the development of sepsis is difficult in humans. We therefore evaluated whether these alterations would appear earlier than global perfusion alterations in a sheep model of peritonitis.


This study was approved by the animal ethical committee of the Université Libre de Bruxelles, and the ARRIVE guidelines for animal research were adhered to (17). Eight young adult female Swifter sheep weighing between 24 and 34 kg and aged between 10 and 12 months were used for this study; seven of these animals were included in a previous experiment (18). Before the experiment, animals were housed in an enclosure for middle-size animals at the Erasmus Campus animal laboratory of the Université Libre de Bruxelles; a maximum of two animals were allowed in the same enclosure. Our laboratory facilities have a controlled temperature of 20–22°C and all animals had light/dark cycles of 12/12 h. Animals were fasted for 24 h before the experiment with free access to water.

Premedication, anesthesia, and ventilation

All experiments were started in the morning. Premedication was administered in the animal enclosure with an intramuscular mixture of 25 mg/kg of ketamine hydrochloride (Ceva Santé Animal, Libourne, France) and 0.3 mg/kg of midazolam (Braun, Melsungen, Germany). The animals were then moved to the experimental room. After supine positioning of the animals on the surgical table, the cephalic vein was cannulated with a peripheral venous 18-gauge catheter (Surflo IV Catheter, Terumo Medical Company, Belgium) and an infusion of Ringer's lactate (Hartmann, Baxter, Lessines, Belgium) was started at 5 mL/kg/h.

Endotracheal intubation was performed under direct laryngoscopy with an 8-mm endotracheal tube (Hi-Contour, Mallinckrodt Medical, Athlone, Ireland) after injection of a bolus of 10 μg/kg of fentanyl citrate (Janssen, Beerse, Belgium) and 0.1 mg/kg of rocuronium bromide (Esmeron, Organon, Oss, The Netherlands). Mechanical ventilation (Servo 300 ventilator; Siemens-Elema, Solna, Sweden) in volume controlled mode was started with a tidal volume of 10 mL/kg, a respiratory rate of 18 breaths/min, a positive end-expiratory pressure (PEEP) of 10 cmH2O, an inspired oxygen fraction (FiO2) of 0.5, inspiratory time to expiratory time ratio of 1:2, and a square-wave flow pattern. Respiratory rate was titrated to maintain the PaCO2 between 30 and 35 mmHg throughout the experiment trying to avoid any possible effect on systemic hemodynamics (24). All other ventilator parameters remained fixed. Continuous intravenous anesthesia was initiated with an infusion of midazolam (0.4–0.8 mg/kg/h), ketamine hydrochloride (4–8 mg/kg/h), and morphine (0.4–0.8 mg/kg/h) adjusting the doses to achieve an adequate level of anesthesia (absence of limb or head movement or of chewing on painful stimulation). Thereafter a continuous infusion of rocuronium (0.2 mg/kg/h) was added to prevent possible movement artifacts. Intravenous anesthesia and muscular blockade were maintained throughout the experiment, using the same doses.

To prevent gastric distension, a 60 cm plastic tube (inner diameter 1.8 cm) was inserted via the esophagus into the stomach to drain its content. To assess urinary output, a 14F Foley catheter (Beiersdorf AG, Hamburg, Germany) was placed in the bladder.

Surgical preparation and hemodynamic monitoring

The right external carotid artery was surgically exposed, a 6F arterial catheter (Vygon, Cirencester, United Kingdom) was introduced and connected to a pressure transducer (Edwards Lifescience, Irvine, Calif) zeroed at mid-chest level and connected to a Siemens SC 9000 monitor (Siemens-Elema, Solna, Sweden) for continuous monitoring of systemic systolic, diastolic, and mean (MAP) arterial pressures. The right external jugular vein was surgically exposed and a 7.5F introducer (Edwards Lifesciences) was used to insert a 7F pulmonary artery (PAC) catheter (Edwards Lifesciences) into the pulmonary artery guided by pressure curves until an adequate pulmonary artery occlusion pressure (PAOP) was obtained. Proximal and distal lumens were connected to different pressure transducers (Edwards Lifesciences) and zeroed at mid-chest level for continuous monitoring of central venous pressure (CVP) and pulmonary pressures. The PAC was connected to a Vigilance I monitor (Edwards Lifesciences) for measurement of cardiac output (CO) by continuous thermodilution.

The right femoral artery and vein were surgically exposed. A perivascular flow meter probe (Transonic, Ithaca, NY) was gently placed around the artery proximally and connected to a TS420 perivascular flow-meter module (Transonic). A 6F catheter (Vygon) was placed inside the right femoral vein to take intermittent blood samples. The left femoral artery was surgically exposed and a 7.5F introducer (Edward Lifesciences) was inserted. A 7F PAC catheter (Edward Lifescience) was then introduced 13 cm to be positioned at the aortic bifurcation as noted in pilot experiments. Adequate positioning was confirmed by the absence of femoral blood flow (fBF) and the appearance of a descending slope on the continuous NIRS StO2 monitoring when the balloon was inflated before baseline measurements. Through a midline laparotomy, we performed a small cecotomy and collected 1.5 g/kg body weight of feces. A purse string suture was used to close the cecotomy wound and the surrounding area was disinfected with iodine solution. The abdominal laparotomy was closed leaving a plastic tube with a large diameter through the abdominal wall for later injection of feces.

Sepsis induction and resuscitation protocol

After the preparatory surgery, we waited 2 h for stabilization of the muscle microdialysis effluents. During this period we checked that MAP, CO, PaO2, and lactate levels remained within normal limits. Baseline measurements were then performed and peritonitis induced by injecting 1.5 g/kg of the autologous feces into the abdominal cavity.

After induction of peritonitis, a bolus of 10 mL/kg of Ringer's lactate was given over 15 min, followed by a continuous infusion started at a rate of 3 mL/kg/h to avoid hypovolemia. Fluid challenges, using 10 mL/kg of Ringer's lactate over 15 min, were performed whenever PAOP decreased, MAP was <60 mmHg, there was an hourly decrease of more than 10% in MAP, arterial lactate increased to >2 mEq/L or urinary output decreased to <0.5 mL/kg/h. If the fluid challenge was associated with an increase in cardiac output of at least 10%, the rate of infusion of Ringer's lactate was increased by 3 mL/kg/h, up to a maximum infusion rate of 12 mL/kg/h (after which no further fluid challenges were performed). If the MAP remained <60 mmHg for more than 2 h with an arterial lactate >2 mEq/L, the rate of fluid infusion was decreased to 4 mL/kg/h until the end of the experiment. No vasopressors or any other vasoactive agents were used during the experiment. The maximal observation period was 30 h. If animals remained alive at this time point, euthanasia was performed by injecting a large dose of potassium chloride under deep anesthesia.

Blood gas analysis and hemodynamic calculations

We calculated the following hemodynamic parameters: cardiac index (CI) as CO/body surface area (BSA), stroke volume index (SVI) as CI/heart rate (HR), systemic vascular resistance index (SVRI) as [MAP–CVP] × 80/CI and left ventricular stroke work index (LVSWI) as SVI × (MAP – PAOP) × 0.0136.

Blood gas analysis, co-oximetry, electrolyte, glucose, and lactate measurements were performed hourly using a Cobas-b 123 point of care system (Roche Diagnostics, Mannheim, Germany) from arterial, mixed venous, and right femoral venous samples. From these samples we measured arterial hemoglobin (Hb) concentrations; arterial, mixed venous, and femoral venous Hb oxygen (O2) saturation (SaO2, SvO2 and SfvO2); and the arterial, mixed venous, and femoral vein CO2 pressures (PaCO2, PVCO2, PvfCO2). We calculated the arterial O2 content (CaO2) as (SaO2 × 1.34 × Hb) + (0.031 × PaO2), venous O2 content (CvO2) as (SvO2 × 1.34 × Hb) + (0.031 × PVO2), O2 delivery (DO2) as CaO2 × CI, O2 consumption (VO2) as (CaO2 – CvO2) × CI, O2 extraction ratio (O2ER) as (DO2–VO2) × 100/DO2 and systemic venoarterial PCO2 difference (PCO2 gap) as PvCO2 – PaCO2. BSA was calculated as 0.085 × (weight in kg [0.67]) (18).

Peripheral and microcirculatory monitoring

The right posterior leg was shaved and the biceps femoris muscle exposed through a small incision in the skin. A microdialysis probe (CMA 20; M Dialysis AB, Stockholm, Sweden) was placed in situ and connected to a microdialysis syringe pump (CMA 402, M Dialysis AB) infusing T1 peripheral liquid (M Dialysis AB) at a rate of 0.3 μL/min. Samples were collected hourly and processed immediately in a microdialysis analyzer (ISCUS, M Dialysis AB) to measure the lactate, pyruvate, and lactate/pyruvate (L/P) ratio concentrations. Muscle StO2 was evaluated using a NIRS spectrometer (InSpectra 650, Hutchinson Technology, Hutchinson, Minn.), with a 15-mm-probe on the external face of the right posterior leg. Arterial VOTs were performed by inflating the intra-aortic PAC balloon for periods of 2 min, followed by rapid deflations. VOTs were performed at baseline (just before induction of peritonitis), and then every 4 h when StO2 values remained stable for at least 1 min. During each VOT we calculated the basal StO2, the total hemoglobin index (THI), the descending (Desc) and ascending (Asc) slopes, the negative (Neg) and positive (Pos) areas (In Spectra Software Analyzer version 3.0, Hutchinson Technology). The ratio of areas (AUC ratio) was calculated as the Pos/Neg area. We calculated the biceps femoris muscle VO2 (NIRS VO2) using the following formula: ((basal THI + minimal THI)/2) × (Desc slope) during the occlusion phase.

To calculate the femoral arterio-venous difference in lactate (fLACT gap), heparinized samples were collected from the arterial and femoral venous catheters at baseline and before each VOT. Lactate concentration was measured in triplicate using a YSI 2300 STAT Plus analyzer (YSI Incorporated Lifesciences, Yellow Springs, Ohio) and the mean values recorded. All blood samples were immediately centrifuged and plasma or serum separated and stored at −80°C before analysis.

We calculated regional variables of O2 transport and consumption in the right posterior leg using the following formulas: the regional femoral venous oxygen content (fCvO2) as (SvfO2 × 1.34 × Hb) + (0.031 × PvfO2); the leg O2 delivery (Leg DO2) as CaO2 × fBF; the leg O2 consumption (Leg VO2) as (CaO2–fCvO2) × fBF; the leg O2 extraction ratio (Leg O2ER) as Leg VO2 × 100/Leg DO2; and the femoral veno-arterial PCO2 difference (fPCO2 gap) as PvfCO2–PaCO2.

Statistical analysis

Statistical analysis was performed using SPSS 19.0 (IBM, New York, NY) software. Data are presented as median values with percentiles 25 and 75 or numbers of cases with percentage, unless otherwise specified. Changes over time for continuous variables were evaluated using a Friedman test. Comparisons with baseline values were performed using a Wilcoxon test. The Spearman rank coefficient (r) was used to evaluate the correlation between different variables correcting for the effect in each studied animal. Correlations were considered as significant only when the P value for the Spearman rank coefficient test was <0.05. The degree of correlation was defined as strong when r was >0.7, moderate when r was between 0.4 and 0.7 and weak when r was <0.4.

We considered a cutoff value for the Asc slope of 120 %/min (=2%/s), selected as being less than the best cutoff value (2.55%/s) identified as discriminating between septic survivors and nonsurvivors in a human prospective study and less than the minimum value reported in a healthy volunteer cohort (10, 19).

To understand the role of MAP on the Asc slope, we calculated relative changes (percentages) from baseline values using the following formulas: relative MAP = (MAP × 100/baseline MAP) and relative Asc slope = (Asc slope × 100/baseline Asc slope). Then, we plotted them together and calculated their correlation using the Spearman rank coefficient (r) and the slope (m) of the linear regression for each animal and for the whole group.

All statistics were two-tailed and a P < 0.05 was considered to be statistically significant.


The systemic hemodynamic and perfusion indexes are presented in Table 1 and Figure 1. There was a significant decrease in MAP and a progressive increase in body temperature over time. There were no significant changes in HR or CI over time, but SVR decreased. Arterial pH and PaO2 decreased, but SvO2, PCO2 gap and arterial lactate levels remained stable. DO2, VO2, and O2ER did not change significantly. The median time to achieve the maximum infusion fluid rate infusion of 12 mL/kg/h was 9 (range 6–10) h. Fluid challenges were performed because of a decrease in PAOP in 19 cases (59%), an hourly MAP decrease of more than 10% in 11 cases (34%), an arterial lactate >2 mEq/L in 1 case (3%) and a urinary output <0.5 mL/kg/h in 1 case (3%). The total volume of crystalloid administered during the experiment was 11.2 (10.7–11.5) mL/kg/h. After 30 h, 4 animals still had a MAP >60 mmHg, 2 animals had an MAP <60 mmHg, and 2 animals had already died (from profound hypotension secondary to refractory septic shock).

Table 1
Table 1:
Systemic variables before each vascular occlusion test (VOT)
Fig. 1
Fig. 1:
Boxplots for mean arterial pressure, systemic vascular resistance index (SVRI), mixed venous oxygen saturation (SvO2), and arterial lactate before each vascular occlusion test (VOT) during the experiment.

The main peripheral variables (including NIRS-derived variables) are shown in Table 2 and Figure 2. Femoral venous pH and SfvO2 decreased significantly from 4 h after sepsis induction, whereas fPCO2 gap and Leg O2ER increased significantly. Muscle NIRS StO2 was significantly lower than at baseline from 8 h after sepsis induction and the StO2 Asc slope from 12 h. Femoral blood flow and muscle pyruvate decreased from 16 h leading to a concomitant increase in the muscle L/P ratio. Leg DO2 was significantly lower than baseline only at 24 and 28 h and Leg VO2 remained at similar values to baseline after a temporary increase at 4 h.

Table 2
Table 2:
Peripheral variables before each vascular occlusion test (VOT)
Fig. 2
Fig. 2:
Boxplots for near-infrared spectroscopy (NIRS) tissue oxygen saturation (StO2), descending slope (Desc slope), ascending slope (Asc slope), and positive area during each vascular occlusion test (VOT) during the experiment.

NIRS StO2 values correlated better with the SvfO2 (r = 0.820) than with the SvO2 (r = 0.436) (Fig. 3). There was a moderate correlation between NIRS THI and arterial hemoglobin (r = 0.507). There was a moderate correlation between the NIRS StO2 Desc slope and the SvO2 (r = 0.432) and a weak correlation between the NIRS StO2 Desc slope and the SvfO2 (r = 0.336). There was a weak correlation between NIRS VO2 and SvO2 (r = 0.358) but a strong correlation between NIRS VO2 and the SvfO2 (r = 0.733).

Fig. 3
Fig. 3:
Correlation of muscle near infrared spectroscopy (NIRS) tissue oxygen saturation (StO2) with mixed venous oxygen saturation (SvO2) and femoral oxygen venous saturation.

There was a moderate correlation between the NIRS StO2 Asc slope and both MAP (r = 0.635) and arterial lactate (r = 0.627) but no correlation with the femoral blood flow pre-VOT (r = 0.205). There was a moderate correlation of the NIRS StO2 Asc slope with muscle lactate (r = −0.447) and a weak correlation with the L/P ratio (r = −0.354). There was a strong correlation of the NIRS Asc slope with the NIRS Pos Area (r = 0.702), but no correlation with the NIRS Des slope (r = 0.294) or NIRS VO2 (r = 0.322). There was a moderate correlation of the femoral blood flow and both the fPCO2 gap (r = −0.527) and the MAP (r = 0.479).

The time-course of the NIRS Asc slope compared with that of arterial lactate and of MAP is shown in Figure 4, using a cutoff value of >60 mmHg for MAP, of <2 mEq/L for arterial lactate, and of >120 %/min for the NIRS Asc slope. Values for the NIRS Asc slope crossed the cutoff earlier than did MAP or arterial lactate values (see also Table 1S in the supplemental digital content,

Fig. 4
Fig. 4:
Simultaneous near infrared spectroscopy (NIRS) ascending slope (Asc slope) and arterial lactate or mean arterial pressure (MAP) measurements throughout the experiment.

Relative changes from baseline (percentages) for NIRS Asc slope as a function of the MAP are shown in Figure 5. The mean slope of individual values (m = 1.2) and the slope for the whole dataset (m = 1.4) confirm that in general the reduction for the Asc slope was larger than the decrease in MAP.

Fig. 5
Fig. 5:
Simultaneous relative values as percentages for near infrared spectroscopy (NIRS) ascending slope (Asc slope) and mean arterial pressure (MAP).

An example of the evolution of NIRS VOT-derived variables as recorded on the NIRS monitor is shown in Figure 6; the severely impaired NIRS VOT shows a biphasic response of the Asc slope.

Fig. 6
Fig. 6:
Evolution of near infrared spectroscopy (NIRS) vascular occlusion test (VOT) from one sheep at baseline (A), after 8 h (B), 20 h (C), and 28 h (D).


This is the first study to evaluate in a well-controlled environment the time-course of NIRS variables including VOT-derived measurements together with other markers of tissue hypoperfusion and global hemodynamics in sepsis. Our animal model of peritonitis was characterized by a progressive decrease in arterial pressure and systemic vascular resistance. The NIRS Asc slope reached clinically relevant abnormal values in most animals from 12 h, whereas MAP, SvO2, and arterial lactate reached clinically relevant abnormal values only toward the end of the experiment.

It is well known that the NIRS Asc slope is decreased in sepsis and that nonsurvivors have persistently lower values than survivors during the ICU stay (10, 16). In our study, all animals had normal Asc slope values at baseline, and the slopes decreased before other systemic variables became abnormal. In the later stages of more severe global hypoperfusion, we observed for the first time a biphasic Asc slope, with a complete absence of the positive reperfusion area, probably explained by a very short postocclusive reactive hyperemia period that ended before the StO2 had returned to pre-VOT values. Other NIRS-derived variables, such as the StO2, also decreased early but the large intersubject variability makes it difficult to determine a useful cutoff value. The Pos area showed an initial increase followed by a progressive decrease until its complete disappearance during septic shock. Finally, the NIRS VOT-derived parameters of local VO2, such as the NIRS Desc slope and NIRS VO2, did not change significantly and were not correlated with the changes in the Asc slope or other global variables. Our regional measurements of leg VO2 using Fick's equation confirm that it did not decrease significantly from baseline values.

Our study clearly demonstrates that NIRS-derived variables correlate better with local peripheral than with global systemic variables. We found that the StO2 correlated better with the SvfO2 than with the SvO2, which may be explained by the redistribution of flow during sepsis. During exercise in healthy volunteers, Mancini et al. (20) showed a good correlation (r = 0.92) between NIRS StO2 values and deep forearm vein oxygen saturation. But McKinley et al. (21) reported no correlation of StO2 with SvO2 in trauma patients (r = 0.55, P = 0.1) and Mesquida et al. (13) noted only a weak correlation of StO2 with ScvO2 in patients with sepsis (r = 0.04). Moreover, Podbregar and Mozina showed a good correlation (r = 0.69) between SvO2 and muscle StO2 in patients with severe heart failure, but this correlation disappeared in patients with concomitant sepsis (r = −0.09) (14). This discordance between global and peripheral measurements has also been described in ICU populations in which SvfO2 was shown to be an unreliable surrogate for ScvO2(22). Indeed, central values reflect the mix of all regional blood flow values; a decrease in blood flow to one region will decrease its contribution to the global flow, but may not be reflected in the overall global value if blood flow is simultaneously increased in other regions.

Of note, recent studies have suggested that the NIRS signal is not only affected by changes in the contents of oxygenated and deoxygenated Hb, but also by oxygenated and deoxygenated forms of myoglobin in the tissue (23). We are not able to distinguish between these proteins in our study, but do not believe this factor would have influenced our conclusions as the proteins share similar functions and are involved in the transient storage and transport of oxygen into the muscle tissue (24).

To preserve blood flow to vital organs like the brain and heart in shock, there is an initial redistribution of peripheral flow. In a dog model of hemorrhagic shock, skeletal muscle PO2 decreased earlier than did MAP (25). In dog models of Escherichia coli infusion, muscle PO2 decreased by 52% already after 30 min, whereas arterial pressure remained unchanged (5). In critically ill patients, Beerthuizen et al. (26) also demonstrated that low skeletal muscle PO2 (<22.5 mmHg) preceded the development of shock. In a series of 36 patients with hemorrhagic shock due to polytrauma, Burša and Pleva showed that L/P ratios determined by microdialysis catheters placed in the deltoid muscle increased 10 h before ScvO2 decreased (27). We also detected significant changes in SvfO2, leg O2ER, fPCO2 gap, femoral venous pH, and NIRS StO2 in the early phases after sepsis induction, which can be explained by the regional decrease in blood flow. Vallet et al. (28), in a model of isolated dog hindlimbs perfused with a pump-membrane oxygenator system, demonstrated that when blood flow was reduced (ischemic hypoxia) there was a concomitant increase in the leg PCO2 gap and O2ER and decrease in the SvfO2. Beilman et al. (29), using a similar design, demonstrated that NIRS StO2 decreased progressively with controlled reductions in regional blood flow. All these findings support the concept that regional perfusion decreases before other alterations appear and that this effect is likely the cause of subsequent alterations in microvascular reactivity.

Our peripheral muscle microdialysis measurements were also affected earlier than systemic perfusion variables. After 16 h there was a marked decrease in muscle pyruvate concentrations, with a slight increase in muscle lactate concentrations, resulting in increased muscle L/P ratios in some animals. Data from muscle microdialysis in sepsis are scarce. In a rat lipopolysaccharide model in which the subcutaneous tissue was monitored using microdialysis, Ohashi et al. (30) observed a significant increase in lactate after 150 min and in the L/P ratio after 300 min. In a pig endotoxemia model, Mutschler et al. (31) detected an increase in the L/P ratio in the quadriceps muscle. In an endotoxic shock pig model, Klaus et al. (32) observed a progressive increase in lactate in the peripheral muscle, but they could not measure pyruvate concentrations. Although all these studies used severe endotoxic models (we used a bacterial septic model) and followed the animals for just a few hours (we collected data for 30 h), all showed the development of cellular hypoxia with anaerobic metabolism in the advanced phases of septic shock.

Finally, we observed that the NIRS Asc slope decreased earlier than the L/P ratios, suggesting that microvascular alterations precede the metabolic shift towards anaerobiosis in the muscle during sepsis. There are few published data with simultaneous measurements of the local microcirculation and metabolism. However, a previous study in the same animal model, showed that different parameters of brain microcirculation (assessed by videomicroscopy) decreased before the brain L/P ratio increased, and this latter variable increased mostly when hypotension appeared (33). In a mouse model of peritonitis by cecal ligation and puncture, Fries et al. (34) showed concomitant decreases in the gastric and buccal microvascular blood flows when systemic lactate started to increase. Nevertheless, a recent study from our group in the same sheep model of bacterial peritonitis, showed that earlier increases in systemic and renal lactate were accompanied by proportional increases in the pyruvate, and that the systemic lactate reached values >2 mmol/L and the L/P ratio increased significantly only late (when hypotension was established) (35), suggesting that initial increases in systemic lactate do not necessarily indicate anaerobic metabolism.

Our study has several limitations. First, the intermittent arterial VOTs used in this septic model can improve hemodynamics and prolong survival time by inducing ischemic preconditioning (18), which may have added a confounding factor. It is obviously impossible to perform a VOT without inducing remote ischemic conditioning (19, 36), but we have shown in healthy volunteers that the repetition of VOTs modifies the Desc slope, but not the Asc slope (19), which was most altered in the present study. Moreover, we studied the sequential order of events and each animal served as its own control. Second, we performed intravascular arterial occlusions (and not external VOTs as performed in human experiments), increasing the risk of thrombotic events. During our pilot studies it was impossible to obtain flow obstruction by external compression because of the rear leg anatomy of the sheep. Importantly, our continuous record of perivascular blood flow excluded any major occlusive vascular event and we recently published results from a similar sheep sepsis model, but without intra-aortic catheters or VOTs, in which the NIRS StO2 and microdialysis data followed similar trends (37). Third, using changes in MAP and arterial lactate, but not in NIRS parameters, as criteria for a fluid challenge may have interfered with our time-to-event analysis. However, most of the fluid challenges were performed because of a decrease in PAOP and none because of an MAP <60 mmHg. Moreover, fluid challenges were only administered during the first 10 h of the experiment and the main findings of our study appeared thereafter, when fluid infusion was already stable. Fourth, we did not include a control group receiving only anesthesia, but we have previously published data from studies that included this type of cohort and showed no significant changes in hemodynamic or metabolic variables over time (35). Finally, the NIRS Asc slope (as a measure of postocclusive vascular reactive hyperemia) relies on local release of different vasodilatory molecules and a driving force that is influenced by the MAP. Both the Asc slope and the MAP decreased progressively during the experiment and were moderately correlated. Additional experiments while maintaining the baseline MAP stable with norepinephrine would be informative, but still controversial as vasoconstrictive agents will compete with the local vasodilation. As patients with sepsis do not usually receive vasoactive agents when initially evaluated in the emergency department and because our data showed that the changes in the Asc slope reached clinically relevant abnormal values earlier than the MAP did, we believe that the NIRS Asc slope (assessed using thenar NIRS with a VOT conducted by inflating a blood pressure cuff around the arm) could be used as an early indicator of sepsis severity; clinical studies are needed to test this hypothesis.


Peripheral muscle NIRS VOT-derived variables are altered earlier than other systemic perfusion variables during sepsis. These data suggest that microvascular reactivity assessed using the NIRS Asc slope can serve as an early warning sign of altered peripheral perfusion in sepsis.


1. Levy MM, Artigas A, Phillips GS, Rhodes A, Beale R, Osborn T, Vincent JL, Townsend S, Lemeshow S, Dellinger RP. Outcomes of the Surviving Sepsis Campaign in intensive care units in the USA and Europe: a prospective cohort study. Lancet Infect Dis 2012; 12:919–924.
2. PRISM Investigators. Early, goal-directed therapy for septic shock—a patient-level meta-analysis. N Engl J Med 2017; 376:2223–2234.
3. Ruokonen E, Takala J, Kari A, Saxen H, Mertsola J, Hansen EJ. Regional blood flow and oxygen transport in septic shock. Crit Care Med 1993; 21:1296–1303.
4. Lam C, Tyml K, Martin C, Sibbald W. Microvascular perfusion is impaired in a rat model of normotensive sepsis. J Clin Invest 1994; 94:2077–2083.
5. Beerthuizen GI, Goris RJ, Kreuzer FJ. Skeletal muscle Po2 during imminent shock. Arch Emerg Med 1989; 6:172–182.
6. Vallet B, Lund N, Curtis SE, Kelly D, Cain SM. Gut and muscle tissue PO2 in endotoxemic dogs during shock and resuscitation. J Appl Physiol (1985) 1994; 76:793–800.
7. De Backer D, Orbegozo Cortes D, Donadello K, Vincent JL. Pathophysiology of microcirculatory dysfunction and the pathogenesis of septic shock. Virulence 2014; 5:73–79.
8. Cohn SM, Nathens AB, Moore FA, Rhee P, Puyana JC, Moore EE, Beilman GJ. Tissue oxygen saturation predicts the development of organ dysfunction during traumatic shock resuscitation. J Trauma 2007; 62:44–54.
9. Doerschug KC, Delsing AS, Schmidt GA, Haynes WG. Impairments in microvascular reactivity are related to organ failure in human sepsis. Am J Physiol Heart Circ Physiol 2007; 293:H1065–H1071.
10. Creteur J, Carollo T, Soldati G, Buchele G, De Backer D, Vincent JL. The prognostic value of muscle StO2 in septic patients. Intensive Care Med 2007; 33:1549–1556.
11. Beilman GJ, Groehler KE, Lazaron V, Ortner JP. Near-infrared spectroscopy measurement of regional tissue oxyhemoglobin saturation during hemorrhagic shock. Shock 1999; 12:196–200.
12. Mulier KE, Skarda DE, Taylor JH, Myers DE, McGraw MK, Gallea BL, Beilman GJ. Near-infrared spectroscopy in patients with severe sepsis: correlation with invasive hemodynamic measurements. Surg Infect (Larchmt) 2008; 9:515–519.
13. Mesquida J, Masip J, Gili G, Artigas A, Baigorri F. Thenar oxygen saturation measured by near infrared spectroscopy as a noninvasive predictor of low central venous oxygen saturation in septic patients. Intensive Care Med 2009; 35:1106–1109.
14. Podbregar M, Mozina H. Skeletal muscle oxygen saturation does not estimate mixed venous oxygen saturation in patients with severe left heart failure and additional severe sepsis or septic shock. Crit Care 2007; 11:R6.
15. Bopp CM, Townsend DK, Barstow TJ. Characterizing near-infrared spectroscopy responses to forearm post-occlusive reactive hyperemia in healthy subjects. Eur J Appl Physiol 2011; 111:2753–2761.
16. Shapiro NI, Arnold R, Sherwin R, O’Connor J, Najarro G, Singh S, Lundy D, Nelson T, Trzeciak SW, Jones AE. The association of near-infrared spectroscopy-derived tissue oxygenation measurements with sepsis syndromes, organ dysfunction and mortality in emergency department patients with sepsis. Crit Care 2011; 15:R223.
17. Kilkenny C, Browne WJ, Cuthill IC, Emerson M, Altman DG. Improving bioscience research reporting: the ARRIVE guidelines for reporting animal research. J Pharmacol Pharmacother 2010; 1:94–99.
18. Orbegozo Cortes D, Su F, Santacruz C, Hosokawa K, Donadello K, Creteur J, De Backer D, Vincent JL. Ischemic conditioning protects the microcirculation, preserves organ function, and prolongs survival in sepsis. Shock 2016; 45:419–427.
19. Orbegozo Cortes D, Puflea F, De Backer D, Creteur J, Vincent JL. Near infrared spectroscopy (NIRS) to assess the effects of local ischemic preconditioning in the muscle of healthy volunteers and critically ill patients. Microvasc Res 2015; 102:25–32.
20. Mancini DM, Bolinger L, Li H, Kendrick K, Chance B, Wilson JR. Validation of near-infrared spectroscopy in humans. J Appl Physiol 1994; 77:2740–2747.
21. McKinley BA, Marvin RG, Cocanour CS, Moore FA. Tissue hemoglobin O2 saturation during resuscitation of traumatic shock monitored using near infrared spectrometry. J Trauma 2000; 48:637–642.
22. van Beest PA, van der Schors A, Liefers H, Coenen LG, Braam RL, Habib N, Braber A, Scheeren TW, Kuiper MA, Spronk PE. Femoral venous oxygen saturation is no surrogate for central venous oxygen saturation. Crit Care Med 2012; 40:3196–3201.
23. Davis ML, Barstow TJ. Estimated contribution of hemoglobin and myoglobin to near infrared spectroscopy. Respir Physiol Neurobiol 2013; 186:180–187.
24. Ordway GA, Garry DJ. Myoglobin: an essential hemoprotein in striated muscle. J Exp Biol 2004; 207:3441–3446.
25. van der Kleij AJ, de Koning J, Beerthuizen G, Goris RJ, Kreuzer F, Kimmich HP. Early detection of hemorrhagic hypovolemia by muscle oxygen pressure assessment: preliminary report. Surgery 1983; 93:518–524.
26. Beerthuizen GI, Goris RJ, Kreuzer FJ. Early detection of shock in critically ill patients by skeletal muscle PO2 assessment. Arch Surg 1989; 124:853–855.
27. Bursa F, Pleva L. Anaerobic metabolism associated with traumatic hemorrhagic shock monitored by microdialysis of muscle tissue is dependent on the levels of hemoglobin and central venous oxygen saturation: a prospective, observational study. Scand J Trauma Resusc Emerg Med 2014; 22:11.
28. Vallet B, Teboul JL, Cain S, Curtis S. Venoarterial CO(2) difference during regional ischemic or hypoxic hypoxia. J Appl Physiol 19852000; 89:1317–1321.
29. Beilman GJ, Myers D, Cerra FB, Lazaron V, Dahms RA, Conroy MJ, Hammer BE. Near-infrared and nuclear magnetic resonance spectroscopic assessment of tissue energetics in an isolated, perfused canine hind limb model of dysoxia. Shock 2001; 15:392–397.
30. Ohashi H, Kawasaki N, Komatsu H, Wada T, Hosoyama A, Hanyu N, Kobayashi K, Ohashi M, Taira Y. Microdialysis detection of lactate in subcutaneous tissue as a reliable indicator of tissue metabolic disorders in an animal sepsis model. J Smooth Muscle Res 2011; 47:37–46.
31. Mutschler DK, Eriksson MB, Wikstrom BG, Lind L, Larsson A, Bergren-Kiiski R, Lagrange A, Nordgren A, Basu S. Microdialysis-evaluated myocardial cyclooxygenase-mediated inflammation and early circulatory depression in porcine endotoxemia. Crit Care Med 2003; 31:1780–1785.
32. Klaus S, Heringlake M, Block K, Nolde J, Staubach K, Bahlmann L. Metabolic changes detected by microdialysis during endotoxin shock and after endotoxin preconditioning. Intensive Care Med 2003; 29:634–641.
33. Taccone FS, Su F, De Deyne C, Abdellhai A, Pierrakos C, He X, Donadello K, Dewitte O, Vincent JL, De Backer D. Sepsis is associated with altered cerebral microcirculation and tissue hypoxia in experimental peritonitis. Crit Care Med 2014; 42:e114–e122.
34. Fries M, Weil MH, Sun S, Huang L, Fang X, Cammarata G, Castillo C, Tang W. Increases in tissue Pco2 during circulatory shock reflect selective decreases in capillary blood flow. Crit Care Med 2006; 34:446–452.
35. Post EH, Su F, Hosokawa K, Taccone FS, Herpain A, Creteur J, Vincent JL, De Backer D. Changes in kidney perfusion and renal cortex metabolism in septic shock: an experimental study. J Surg Res 2017; 207:145–154.
36. Kharbanda RK, Mortensen UM, White PA, Kristiansen SB, Schmidt MR, Hoschtitzky JA, Vogel M, Sorensen K, Redington AN, MacAllister R. Transient limb ischemia induces remote ischemic preconditioning in vivo. Circulation 2002; 106:2881–2883.
37. Orbegozo D, Su F, Santacruz C, He X, Hosokawa K, Creteur J, De Backer D, Vincent JL. Effects of different crystalloid solutions on hemodynamics, peripheral perfusion, and the microcirculation in experimental abdominal sepsis. Anesthesiology 2016; 125:744–754.

CO2 gap; microcirculation; microdialysis; near-infrared spectroscopy; noninvasive monitoring; oxygen consumption; perfusion; sepsis; vascular reactivity

Supplemental Digital Content

Copyright © 2017 by the Shock Society