Materials and Methods

Animals—Fourteen healthy purpose-bred university-owned adult Beagles were enrolled in the study. Dogs were determined to be healthy on the basis of results of physical examination, CBC, and serum biochemical profile. Food was withheld for 12 hours prior to each experiment, but dogs had unrestricted access to water. Dogs were housed and cared for according to the Association for Assessment and Accreditation of Laboratory Animal Care adopted guidelines. The Virginia Tech Animal Care and Use Committee approved this study.

Instrumentation—Following preoxygenation with 100% oxygen, anesthesia was induced with 5% isoflurane^a delivered by facemask. An endotracheal tube was placed and connected to a circle breathing system. Dogs were positioned in dorsal recumbency, and anesthesia was maintained with 1.3% to 2% isoflurane in 100% oxygen during the instrumentation period. The cephalic vein and the right dorsal pedal artery were catheterized with 20-gauge 32-mm-long over-the-needle catheters. A 7F 20-cm double lumen catheter^b was placed in the left jugular vein, and a 5F 8.5-cm hemostatic introducer^c was placed in the right jugular vein. Dogs were then positioned in left lateral recumbency, and a thermodilution balloon catheter^d with heparin coating was advanced to the pulmonary artery. Placement was confirmed by observation of the characteristic waveform.

Dogs were instrumented for continuous monitoring of respiratory rate, Petco₂, bispectral index, HR, SAP, MAP, diastolic arterial blood pressure, CVP, and PAP.^e Pressure transducers were calibrated with a calibration device,^f then placed and zeroed at the level of the sternum. Bispectral index was continuously acquired and displayed by use of a proprietary bispectral index module in the monitor.^e Stainless steel subdermal needle electrodes with a 3-lead reference montage were connected to a proprietary sensor^g and arranged as described.¹³ Briefly, the primary lead was placed on the midline over the frontal bone at the level of the zygomatic processes, the ground lead was placed rostral to the base of the left ear, and the secondary lead was placed over the temporal bone. Esophageal temperatures were continuously monitored^e and maintained between 36° and 38°C by use of a forced-air patient warmer.^h Arterial and mixed venous blood samples were obtained and immediately analyzedⁱ for Hb, Pao₂, pH, Po₂, BE, and lactate concentration. Cardiac output measurements were obtained in triplicate by thermodilution by use of 3 mL of iced 5% dextrose injected into the proximal port of the Swan-Ganz catheter. Cardiac output was recorded as the mean of 3 measured values with < 10% variation. Pulmonary artery wedge pressure was measured after CO determination with the Swan-Ganz catheter. Cardiac index, SVI, SVRI, Cao₂, Co₂, o₂, and Eo₂ were calculated with standard formulas.¹⁴ Tissue oxygen saturation was determined with a Sto₂ monitor.^j The Sto₂ probe was placed over the sartorius muscle of the left pelvic limb (dependent limb).^j Prior to placement of the Sto₂ probe, the site was shaved to minimize interference from the hair during signal acquisition. Adequate signal strength was determined by detection of a THI > 5. Tissue Hb index is a measure of Hb in the monitored tissue and an indicator of signal strength. In human patients, a THI > 5 indicates sufficient Hb to obtain an accurate Sto₂ reading. The Sto₂ was continuously monitored at 2-second intervals over 1 minute with the Sto₂ monitor. The Sto₂ recorded was the mean determined over a designated minute with a THI > 5.

Experimental protocol—Following instrumentation, anesthesia was transitioned from isoflurane to a constant rate infusion of propofol,^a as part of an unrelated study, to facilitate mechanical ventilation with a specific ventilator. The constant rate infusion of propofol was adjusted to maintain anesthesia and a bispectral index < 65. Dogs were paralyzed with a loading dose of rocuronium bromide^k (0.2 mg/kg, IV), followed by a constant rate infusion (0.8 mg/kg/h) to facilitate ventilation. Mechanical ventilation^l with 100% oxygen was continued to maintain normocapnia (Petco₂, 35 to 45 mm Hg). Before proceeding with the experimental protocol, complete washout of isoflurane was confirmed by use of the gas analyzer.^e Dogs were evaluated at 3 volumetric states: euvolemia, hypovolemia, and hypervolemia. After a 30-minute stabilization period, euvolemic data were collected (baseline). Following this, a hypovolemic state was induced by removing blood through the double lumen catheter in the left jugular vein, until the MAP stabilized at 40 ± 5 mm Hg. The blood bag collection set, containing a citrate-phosphate-dextrose-adenine solution, was suspended from a hook on the top of a vacuum chamber placed on a scale to measure total blood drawn. In preparation for autotransfusion, blood was stored in an incubator at 37°C. Once the target MAP was achieved, dogs were maintained at a steady hemodynamic plane for a minimum of 10 minutes before data were collected (hypovolemia T1). If MAP increased to greater than the target value, additional blood was removed until the target MAP was restored. Following restoration of the target MAP, the 10-minute equilibration period was restarted prior to data collection. Dogs were maintained in a hypovolemic state for 20 minutes, and data were collected prior to resuscitation with the shed blood (hypovolemia T2). Shed blood was administered over 20 minutes, and data were collected immediately after to confirm restoration of euvolemia (after transfusion). Then hetastarch^a was administered at 20 mL/kg over 20 minutes to achieve hypervolemia and immediately afterward data were collected (hypervolemia). Dogs were allowed to recover. A dose of buprenorphine^m (0.02 mg/kg, IV) was administered during recovery. Dogs were monitored after surgery until they were walking without ataxia. If signs of hemorrhage were detected at the catheter sites, the dogs were transferred to the intensive care unit for 12 hours. The following day, a physical examination was performed, and PCV, total solids concentration, and BUN were checked. All dogs survived.

Statistical analysis—Main outcomes of interest for this study were Sto₂ and Do₂. Other outcomes included respiratory rate, Petco₂, HR, MAP, CVP, PAP, temperature (esophageal probe), Hb concentration, Pao₂, Pao₂, pH, Po₂, BE, and lactate concentration. Normal probability plots revealed that all outcomes followed a normal distribution. Subsequently, data were summarized as mean ± SD values. Effects of time period (normovolemia [baseline], hypovolemia with target MAP < 40 mm Hg achieved and maintained steady for 10 minutes [hypovolemia T1], then 20 minutes later [hypovolemia T2], following resuscitation with shed blood [after transfusion] and after 20 mL of hetastarch/kg [hypervolemia]) on each outcome were assessed by use of mixed-model ANOVA followed by the Tukey procedure for multiple comparisons. The linear model specified time period as a fixed effect and dog as the random effect. To test the association between Do₂ and Sto₂, dog-level data for Do₂ and Sto₂ were initially summarized (by use of a mean value) at each time period before generating the Pearson correlation coefficient. Significance was set at P < 0.05. All analyses were performed with statistical software.ⁿ

Results

All 14 dogs (7 males and 7 females) used in the study recovered from anesthesia without complications. They were all 2 years of age and older, with a mean ± SD body weight of 10 ± 0.72 kg. Cardiopulmonary values were determined (Tables 1–3).

Table 2—

Mean ± SD arterial and mixed venous blood gas values measured in the same dogs as in Table 1.

Variable	Baseline	Hypovolemia T1	Hypovolemia T2	After transfusion	Hypervolemia
Pao2 (mm Hg)	475 ± 39	452 ± 49	450 ± 51	482 ± 36	471 ± 36
BE (mmol/L)	−2.9 ± 2.18	−6.6 ± 2.06*	−8 ± 2.5*†	−7.3 ± 1.73*	−5.2 ± 1.72*†‡§
Po2 (mm Hg)	75 ± 21.42	37 ± 6.96*	43 ± 6.78*	84 ± 15.44†‡	90 ± 21.10*†‡
Hb (g/dL)	11 ± 0.97	10.74 ± 1.59	9.08 ± 1.34*†	10.32 ± 0.87†	8.16 ± 1.29*†§
Lactate (mmol/L)	0.77 ± 0.52	1.06 ± 0.69	1.24 ± 0.76*	1.08 ± 0.56	0.51 ± 0.30†‡§

See Table 1 for key.

Table 3—

Mean ± SD values for cardiopulmonary variables calculated in the same dogs as in Table 1.

Variable	Baseline	Hypovolemia T1	Hypovolemia T2	After transfusion	Hypervolemia
Cao2 (mL/dL)	16.17 ± 1.33	15.74 ± 2.17	13.52 ± 1.84*	15.23 ± 1.14‡	12.35 ± 1.69*†§
Co2 (mL/dL)	13.70 ± 1.48	8.29 ± 2.25*	7.83 ± 1.95*	12.97 ± 1.29†‡	10.54 ± 1.62*†‡§
o2I (mL/min/m2)	89.08 ± 29.79	94.43 ± 41.49	84.22 ± 17.30	104.56 ± 21.64	98.75 ± 21.59
Eo2 (%)	0.15 ± 0.04	0.47 ± 0.12*	0.42 ± 0.10*	0.15 ± 0.03†‡	0.15 ± 0.03†‡
THI	16 ± 2.91	14 ± 2.18*	14 ± 2.40*	16 ± 2.34†‡	15 ± 2.90†‡

See Table 1 for key.

Mean ± SD duration for complete isoflurane washout was 50 ± 11 minutes. Dosage for the constant rate infusion of propofol ranged from 18 to 42 mg/kg/h. Mean time required to achieve the target MAP during hypovolemia was 26.7 ± 15.3 minutes. Blood volume removed from the dogs ranged from 297 to 519 mL, with a mean of 393.6 ± 62.3 mL. The 10-minute stabilization period had to be repeated in 6 dogs following the removal of additional blood. Following initiation of hypovolemia, there was a significant increase in HR, BE, and Eo₂, compared with baseline values. These changes were accompanied by a significant decrease in MAP, CVP, CI, SVI, Po₂, and Cmo₂ (Tables 1–3). From baseline to hypovolemia T1, there was a 67% reduction in Do₂I (619 ± 257 mL/min/m² vs 205 ± 76 mL/min/m²; P < 0.001) and a 17% reduction in Sto₂ (94 ± 4.4% vs 78 ± 12.2%; P < 0.001; Figure 1).

View Full Size

Mean ± SD Do₂I (squares) and Sto₂ (diamonds) values at euvolemia (baseline), after hemorrhage (Hypovolemia T1), 20 minutes later (Hypovolemia T2), after transfusion (posttransfusion), and after hetastarch administration (Hypervolemia) in 14 dogs. *Significantly (P < 0.05) different from baseline. †Significantly (P < 0.05) different from hypovolemia T1. ‡Significantly (P < 0.05) different from hypovolemia T2.

Citation: American Journal of Veterinary Research 75, 1; 10.2460/ajvr.75.1.48

• Download Figure
• Download figure as PowerPoint slide

Mean ± SD duration of the hypovolemic period, from hypovolemia T1 to hypovolemia T2, was 37.8 ± 10.9 minutes. During the hypovolemic period, there was a decrease in BE, SVRI, Hb, and Cao₂ from hypovolemia T1 to T2 (Tables 1–3). There was no significant change in HR, MAP, CI, SVI, Po₂, lactate concentration, and Co₂. However, there was a significant increase in lactate concentration at hypovolemia T2, compared with the euvolemic value (Table 2).

Dogs received a transfusion during a mean of 20 ± 11 minutes. Upon completion of autotransfusion of shed blood, MAP, CVP, CI, SVI, Pvo₂, Hb, lactate concentration, Co₂, and Eo₂ returned to euvolemic values (Tables 1–3). Heart rate, BE, and SVRI failed to return to euvolemic values. In comparison to hypovolemia T2 values, Sto₂ (80 ± 8.5% vs 92 ± 6.45%; P < 0.001) and Do₂I (211 ± 73 mL/min/m² vs 717 ± 221 mL/min/m²; P < 0.001) significantly increased, returning to euvolemic values (Figure 1).

Subsequent to the administration of hetastarch and induction of hypervolemia, MAP, CVP, HR, CI, and Po₂ were significantly increased, compared with euvolemic values (Tables 1 and 2). There was a significant decrease in SVRI, Hb, BE, Cao₂, and Co₂, compared with euvolemic values. No significant differences were observed with respect to SVI, lactate concentration, Do₂I, Sto₂, or the Eo₂, compared with euvolemic values. Across all time points in the study, there was no significant change in Pao₂ or o₂. Comparison of mean Do₂I and Sto₂ across all time points yielded a strong relationship, with a correlation coefficient of 0.97 (P = 0.005).

Discussion

The present study was performed to evaluate Sto₂ by use of NIRS in experimental acute hemorrhagic shock and resuscitation in dogs. We sought to test the hypothesis that tissue-specific oxygenation as measured by Sto₂ would correlate with whole organism Do₂ during hemorrhagic shock and reperfusion.

Following induction of hemorrhagic shock, there was a significant decrease in Do₂I and Sto₂. Although the decrease in Do₂I was greater than the decrease in Sto₂, the data revealed a strong correlation between St and Do₂I during the period between baseline and hypovolemia T1. As dogs received the autotransfusion, leading to the restoration of euvolemia, Sto₂ and Do₂I significantly increased toward baseline values and a strong correlation was again observed between Sto₂ and Do₂I. These results were similar to those obtained in previous studies in other species. McKinley et al¹⁰ monitored Sto₂ during resuscitation and for the first 24 hours following admission of human patients who had sustained severe abdominal trauma; Sto₂ correlated with changes in Do₂ and was useful in guiding restoration of euvolemia. In a study¹⁵ of the relationship between Do₂ and Sto₂ during experimental acute hemorrhagic shock in splenectomized pigs, a target blood volume of 28% was removed. Similar to the present study, changes in Sto₂ induced by hemorrhage were not as profound as those observed in Do₂ and a strong correlation was evident between changes in Sto₂ and Do₂ during hemorrhage and reperfusion. The larger fluctuation in Do₂, compared with that in Sto₂, may be attributable to a buffering effect in the Do₂. There may be a critical point beyond which changes in Do₂ have little effect on St and before which changes in Do₂ significantly affect Sto₂, similar to the oxygen-Hb dissociation curve.

Changes in BE and lactate concentration, the traditional markers of perfusion, lagged behind changes in Do₂I and Sto₂ following induction of hemorrhagic shock and were slower to return to baseline following fluid resuscitation. This finding was in accordance with clinical and laboratory studies,^6,15 in which changes in Sto₂ precede changes in lactate concentration and base deficit. Following induction of hemorrhagic shock, statistical analysis indicated that the increase in lactate concentration was significant, but the increase was not of sufficient magnitude to be considered clinically important. This can be explained by the fact that Eo₂ was increased in an attempt to maintain aerobic metabolism and the body reserves were not exhausted; therefore, critical Do₂ was not achieved. Additionally, o₂ did not change linearly with Do₂, as would be expected after an anaerobic threshold is reached. Conversely, the Sto₂ monitor detected changes in regional tissue perfusion before changes in lactate concentration and BE occurred, suggesting that Sto₂ may be a valuable tool in identifying patients in the initial stage of shock before critical changes in cellular metabolism occur.

Presently, the effects of volume overload on NIRS-acquired Sto₂ are unknown. It has been speculated that volume overload may alter the reflectance by creating scatter, leading to falsely decreased measurements of Sto₂.¹⁶ In the present study, hypervolemia had no effect on Sto₂ or Do₂. However, the dogs in this study were not subject to extreme fluid overload and time was not allowed for fluid shift to occur. As fluid moves into the interstitial space, altering tissue density, the potential exists to affect the reflectance by creating scatter or decreasing tissue penetration by the near-infrared light. The data obtained in the present study indicated that Sto₂ would not be an effective tool for assessing fluid overload. Additional studies are indicated to determine how fluid overload affects Sto₂ over time.

Tissue hemoglobin index is an indicator of signal strength reported by the Sto₂ monitor and is reported as a number in the range of 1 to 99. The THI corresponds to the amount of Hb present in the monitored tissue. A THI > 5.0 indicates sufficient Hb for the monitor to reliably measure Sto₂ in human patients under most circumstances.¹⁷ The relationship between THI and signal accuracy has not been verified in dogs, and on the basis of clinical experience, we believe THI needs to be > 5.0 to indicate a reliable signal. Tissue hemoglobin indices in this study were all > 5.0. The usefulness of THI is still under investigation, but in the present study, changes in THI were similar to those in Do₂ and Sto₂ across all volemic states. These changes were probably caused by extravascular and vascular Hb concentrations resulting from changes in the volume of the underlying vascular bed, as occur during hemorrhage and reperfusion.

It is important to distinguish the difference between Sto₂ and another commonly used oxygen measurement, Spo₂. Whereas pulse oximetry, measures oxygen saturation in arterial blood, Sto₂ is a measure of oxygen saturation in the vascular bed of the muscle. The Sto₂ signal is composed of not only the arterial oxygen but also the oxygen in the capillaries and venous network. Unlike Spo₂ which reflects changes in oxygenation, Sto₂ reflects changes in oxygen extraction at the tissue level.¹⁷ Therefore, when local tissue conditions of supply and consumption change, Sto₂ will change but Spo₂ will not. This would be a typical finding in a patient with adequate ventilation but poor oxygen delivery from causes such as poor CO, anemia, or hypovolemia. In the present study, there were no changes in Spo₂ with changes in volemic states (data not shown).

Problems associated with placement of the NIRS Sto₂ probe and signal acquisition have been reported in human patients. Darker skin pigmentation, areas with large amounts of fat deposition, and interstitial edema have all been implicated in poor signal acquisition.^18–20 In humans, the NIRS Sto₂ probe is placed over the deltoid muscle or more commonly the thenar eminence (the group of muscles on the palm at the base of the thumb). Use of these regions yields a reliable measurement of St as determined by use of the THI. The decision to place the NIRS Sto₂ probe on the medial aspect of the thigh over the sartorius muscle in the present study was based on previous work by Hall et al.²¹ In that study,²¹ probe placement was evaluated at 4 areas (lower region of the back, forearm, lateral aspect of the thigh, and medial aspect of the thigh) and it was determined that the medial aspect of the thigh provided the most reliable data. In the present study, the dogs did not have a pigmented area in the thigh and had minimal fat deposition. However, the authors have had difficulty with signal acquisition in the clinical setting in patients with dark hair and pigmented skin, suggesting that melanin may interfere. In the present study, the major challenge was to secure the probe to the skin and still allow good contact. We found that using surgical skin staples around the body of the probe worked well for this purpose.

The present study had several limitations. Blood was removed through an intravascular catheter instead of creating a surgical lesion. Although this methodology allows for standardization among animals, it does not completely reflect the clinical situation and this factor has to be considered when interpreting the results. Furthermore, the study was conducted while dogs received general anesthesia with a fraction of inspired oxygen of 0.95, and resuscitation of trauma patients in the emergency room may occur at lower oxygen concentrations, leading to an inability of Do₂ to meet oxygen demands and therefore a faster onset of a state of shock.

Additionally, these dogs were not splenectomized, which accounted for the variability in the time to achieve target MAP and total blood volume removed. Six of the dogs required removal of additional blood after compensation, which was characterized by tachycardia and an increase in MAP. Despite the variability attributed to the splenic contraction in these dogs, this condition more closely resembled a clinical scenario and allowed for better extrapolation of the data.

Despite NIRS becoming more commonly accepted in human medicine, several obstacles to the use of NIRS for the determination of Sto₂ must be overcome before this technology can gain widespread clinical application in the veterinary field. The technology is cost-prohibitive for the veterinary profession, with the cost of a single-use, disposable NIRS Sto₂ probe at approximately $200. There is considerable debate on how best to use this technology. Presently, we do not know what specific Sto₂ would warrant interventions and what specific interventions should be performed when Sto₂ is low. However, under the experimental conditions of hemorrhagic shock and resuscitation in the study reported here, there was a strong correlation between Sto₂ and Do₂I, suggesting that Sto₂ may be used to estimate Do₂. Additionally, changes in Sto₂ preceded those in BE and lactate concentration, indicating that Sto₂ might be used as an early indicator of hypoperfusion.