In individuals that are anesthetized and mechanically ventilated, pulse pressure changes during the respiratory cycle reflect beat-to-beat variations in cardiac stroke volume caused by mechanical ventilation. Pulse pressure variation is a parameter derived from the direct arterial pressure signal and represents the percentage variation between the PPmax, which is usually recorded during the inspiratory cycle of mechanical ventilation, and the PPmin, which is recorded between tidal breaths, and is calculated as PPV = ([PPmax – PPmin]/[PPmax + PPmin])/2 × 100.12
Similarly, the PVI provides an automated measurement of the percentage variation in the amplitude of the plethysmographic waveform in mechanically ventilated individuals. It is calculated on the basis of the percentage variation in the PI, which is a variable derived from changes in the absorption of light induced by modifications in the blood flow (and stroke volume) to the placement site of a pulse oximeter probe.3,4 The PI is related to the quality of the plethysmographic signal, where PImax and PImin are analogous to the PPmax and PPmin recorded from the direct arterial blood pressure tracing. The PVI = (PImax – PImin)/PImax × 100.
Cyclic oscillations in stroke volume induced by mechanical ventilation are associated with cyclic oscillations in the amplitude of pulse pressure recorded via an intra-arterial catheter, and in the amplitude of the plethysmographic waveform recorded by pulse oximetry.1,3 Because oscillations in stroke volume, PPV, and PVI are increased from reference intervals for patients in which the heart is operating on the preload-dependent limb of the Frank-Starling curve, both parameters can be used to predict fluid responsiveness (ie, the ability to improve stroke volume or cardiac output in response to fluid administration).1–3
In humans, PPVs > 12% to 13% are predictive of fluid responsiveness (ie, volume loading is expected to result in a substantial [> 15%] increase in stroke volume).1,5 Results of 2 studies3,4 indicate that there is a significant linear relationship between PPV and PVI. Because the specificity and sensitivity of PVI to predict which individuals are likely to respond to fluid therapy or volume loading are equivalent to the specificity and sensitivity of PPV to predict the same in critically ill human patients,6 interest in the use of PVI, which is derived from pulse oximetry, to monitor fluid responsiveness in veterinary patients is increasing. In humans, PVI threshold values for identification of fluid responsiveness range from > 10% to > 14% in adult patients and from > 13% to > 18% in pediatric patients.7 The compliance of the chest wall relative to lung compliance has a significant influence on the oscillations in stroke volume induced by mechanical ventilation.8,9 Because the chest wall-to-lung compliance ratio is substantially larger in dogs than in humans,8–10 PPV and PVI values that reflect fluid responsiveness in humans cannot be extrapolated to dogs.
In a study11 conducted by our laboratory group, the PPV increased to > 12% after removal of 30% of the estimated blood volume in 4 of 8 isoflurane-anesthetized dogs. The PPV of those 4 dogs returned to prehemorrhage values following VR with autologous blood, which suggests that, in dogs, a PPV > 12% might be indicative of fluid responsiveness11; however, an actual PPV threshold value has yet to be determined for that species. In that study,11 only the dogs that became hypotensive (MAP < 60 mm Hg) after 30% of the estimated blood volume was removed had an increase in PPV, which suggests that blood loss is associated with substantial increases in PPV only in hypotensive hypovolemic dogs and, in an experimental setting, VR will return the PPV to prehemorrhage values.
The purpose of the study reported here was to compare changes in PPV and PVI induced by hemorrhage followed by VR in isoflurane-anesthetized dogs. Our hypotheses were that hemorrhage sufficient to cause hypotension (MAP < 50 mg Hg) will cause a significant increase in PPV in isoflurane-anesthetized dogs, and VR following that hemorrhage will cause the PPV to return to a value similar to that before hemorrhage. We also hypothesized that the PPV and PVI in isoflurane-anesthetized dogs will change in a similar manner during hypotension-inducing hemorrhage and subsequent VR.
Materials and Methods
Animals
Seven purpose-bred English Pointers with ages ranging from 20 to 23 months and body weights ranging from 19.8 to 30.0 kg were used in the study. All dogs were considered healthy on the basis of results of a physical examination, CBC, serum biochemical analysis, and venous blood gas and electrolyte analysis. All study procedures were reviewed and approved by the Institutional Animal Care Committee of São Paulo State University.
Anesthesia and instrumentation of dogs
For each dog, food but not water was withheld for 12 hours prior to induction of anesthesia. Anesthesia was induced with 5% isofluranea in oxygen delivered via a face mask with a circle breathing circuit until orotracheal intubation could be achieved. After intubation, the endotracheal tube was connected to an anesthesia machine,b and the dog was positioned in lateral recumbency. End-tidal carbon dioxide concentration and ETISO were continuously recorded by an infrared gas analyzer that was integrated into the anesthesia machine. Then, volumecontrolled ventilation was initiated with a fixed tidal volume of 12 mL/kg and inspiration-to-expiration ratio of 1:1.5. Respiratory rate was adjusted as necessary to maintain an end-tidal carbon dioxide concentration of approximately 45 mm Hg, and the vaporized isoflurane concentration was adjusted to maintain the dog at a moderate depth of anesthesia during catheter placement. A forced-air warming devicec was used to maintain the dog's core temperature between 37° and 38°C. A 20-gauge catheter was placed in a cephalic vein for administration of Ringer solution at a rate of 2 mL/kg/h by means of a peristaltic pumpd for the remaining duration of anesthesia and for transfusion of autologous blood during VR.
A 20-gauge catheter was placed in a dorsal pedal artery for collection of arterial blood samples for measurement of Hct, total plasma protein concentration, and temperature-corrected blood gas analysise and for measurement of MAP by use of a fluid-filled pressure transducerf that was connected to the screen of patient monitor.g Pulse pressure variation was derived from the direct blood pressure signal by means of proprietary PPV analysis software,h which synchronizes the PPV calculation recorded from the direct blood pressure signal with the inspiratory and expiratory phases of mechanical ventilation recorded from the airway pressure tracings.i This software automatically rejects the display of PPV in the presence of arrhythmias or airway pressure tracing irregularities because those conditions are associated with erroneous PPV values.1,2 Another 20-gauge catheter was placed in the ipsilateral dorsal pedal artery for withdrawal of blood during controlled hemorrhage.
A clip-type pulse oximeter probej was positioned on the mid portion of the tongue and connected to a pulse oximeter monitork that displayed the PVI and calculated the mean PI. To optimize the PI signal and obtain reliable PVI data, the tongue was gently massaged and the probe was repositioned approximately 3 minutes before data collection at each predetermined time point.
An 8F catheter introducer was placed in a jugular vein through which a 7F thermodilution catheterl was inserted into the pulmonary artery. Correct placement of the thermodilution catheter was determined on the basis of observation of characteristic pressure waveforms on the screen of a second patient monitor.m The CVP and PAOP were recorded from the proximal and distal ports of the thermodilution catheter by means of 2 fluid-filled pressure transducers.f The balloon located at the tip of the thermodilution catheter was insufflated with 0.8 mL of air during measurement of PAOP. Cardiac output was measured by a thermodilution technique in which 5-mL boluses of an ice-cold (3° to 5°C) 5% dextrose solution were injected into the central venous port of the thermodilution catheter. At each data collection time, the cardiac output was calculated as the mean of 3 consecutive measurements.1 The temperature of the injected dextrose solution was measured by an in-line thermistor that was located between the syringe and the proximal injection port of the thermodilution catheter.
Body surface area (m2) was calculated as follows: (weight in grams)2/3 × 10.1 × 10−4. Hemodynamic variables including CI, SI, and SVRI were calculated in accordance with standard formulas.12
Experimental protocol
For each dog after completion of instrumentation, a neuromuscular blockade was induced by administration of atracurium (0.2 mg/kg, IV bolus followed by 0.2 mg/kg/h as a constant rate IV infusion) with a syringe pump.n The effectiveness of the neuromuscular blockade was monitored by a supramaximal train-of-four electrical stimuluso of an ulnar nerve. During the atracurium infusion, the tidal volume was maintained at 12 mL/kg and the inspiration-to-expiration ratio was maintained at 1:1.5; however, the respiratory rate was adjusted to maintain Paco2 at approximately 45 mm Hg while not exceeding a maximum limit of 26 breaths/min. The ETISO was adjusted to maintain MAP at 65 mm Hg (target range, 60 to 70 mm Hg). After the MAP had stabilized within the target range for 10 minutes, dogs with a PPV > 15% were administered a 6% tetrastarch solutionp (5 mL/kg, IV) in an attempt to achieve a PPV < 15%. After the MAP had equilibrated within the target range for at least 30 minutes, baseline cardiopulmonary data were obtained. The ETISO used during collection of baseline data was maintained for the duration of the observation period. Reported ETISO values were corrected to the barometric pressure at sea level (760 mm Hg) by use of the following formula: (barometric pressure of location/760 mm Hg) × obtained ETISO.
Following collection of baseline data, progressive hemorrhage was initiated by withdrawing blood from the catheter in the dorsal pedal artery that was designated for that purpose at a rate of 1 mL/kg/min. To achieve and maintain that rate, the blood volume withdrawn every minute was split into two 20-mL syringes that had been prefilled with an adequate amount of a solution containing sodium citrate, citric acid, dextrose, and adenine to prevent coagulation of the blood volume being withdrawn. The blood was then transferred from the syringes into collection bagsq that did not contain any anticoagulant through a 3-way stopcock system. That process (controlled hemorrhage) was repeated until the MAP decreased to 45 mm Hg (target range, 4 to 50 mm Hg), after which cardiopulmonary data were immediately recorded. Following data collection, VR was achieved by reinfusion of the autologous blood removed during controlled hemorrhage by use of a peristaltic infusion pumpd over a period of 30 minutes. Cardiopulmonary data were recorded again immediately following VR.
Anesthesia recovery
Following data collection after VR, all catheters were removed. The neuromuscular blockade was reversed by administration of neostigmine (0.04 mg/kg, IV) and atropine (0.02 mg/kg, IV and 0.02 mg/kg, IM) once train-of-four stimulation resulted in the spontaneous return of 3 of 4 thoracic limb contractions. Following reversal of the neuromuscular blockade, mechanical ventilation and isoflurane administration were continued until all 4 thoracic limb contractions had returned to baseline conditions and then discontinued. Blood oxygen saturation with the dog breathing room air was monitored by pulse oximetry to ensure that it was adequate (≥ 95%) before the endotracheal tube was removed. A dog was extubated as soon as its swallowing reflex returned. The total anesthesia time (time between endotracheal intubation and discontinuation of isoflurane administration) and recovery time (time between discontinuation of isoflurane administration and achievement of standing without ataxia) were recorded. Each dog was administered meloxicam (0.2 mg/kg, IV) once immediately following extubation for analgesia.
Statistical analysis
The data for each physiologic variable at each collection time (baseline and immediately after controlled hemorrhage and VR) were determined to be normally distributed by means of a Shapiro-Wilk test. For each physiologic variable, a 1-way ANOVA for repeated measures with a post hoc Tukey multiple comparison test was used to compare results among time points. At each time, the association between PPV and PVI was evaluated by means of Pearson least squares linear regression analysis. Results for each variable were reported as the mean ± SD unless otherwise specified, and values of P < 0.05 were considered significant for all analyses. All analyses were performed with a commercially available statistical software program.r
Results
Dogs
For all 7 dogs, the mean ± SD ETISO corrected to sea level required to maintain MAP between 60 and 70 mm Hg was 1.74 ± 0.48%. Prior to collection of baseline (immediately prior to controlled hemorrhage) data, 3 of 7 dogs had a PPV ≥ 15% (range, 17% to 19%) and were administered tetrastarch; after colloid administration, the PPV ranged from 10% to 12% in those dogs. Baseline data were recorded at a mean ± SD of 159 ± 33 minutes after induction of anesthesia. The mean ± SD volume of blood withdrawn during controlled hemorrhage was 25 ± 5 mL/kg, which represented 31 ± 6% of the estimated total blood volume (80 mL/kg) for each dog. The mean ± SD duration of anesthesia was 289 ± 36 minutes. All dogs recovered from anesthesia without complications and had an end-tidal carbon dioxide concentration < 45 mm Hg and blood oxygen saturation (as determined by pulse oximetry) ≥ 95% while breathing room air at the time of extubation. The mean ± SD recovery time (duration from discontinuation of isoflurane administration until the dog achieved a standing position) was 69 ± 59 minutes.
Physiologic variables
Physiologic variables at baseline and immediately following controlled hemorrhage and VR were summarized (Table 1). The mean HR immediately following controlled hemorrhage was significantly decreased from baseline, whereas mean HR recorded after VR was increased from baseline and from values recorded immediately after controlled hemorrhage. Mean SI and CI were decreased from baseline immediately after controlled hemorrhage, whereas after VR, mean SI and CI values were significantly increased from those after controlled hemorrhage but were not significantly different from those at baseline. Mean CVP and PAOP were both decreased from baseline immediately after controlled hemorrhage. Immediately after VR, mean CVP and PAOP were both significantly increased from those after controlled hemorrhage and from baseline. As expected, the MAP was significantly decreased from baseline immediately after controlled hemorrhage; however, although VR increased the MAP from values recorded after controlled hemorrhage, it remained lower than baseline after the end of VR. The mean SVRI was increased from baseline immediately after controlled hemorrhage, whereas VR decreased the mean SVRI from that after controlled hemorrhage and baseline. The mean arterial blood pH and total plasma protein concentration were significantly decreased from baseline immediately after controlled hemorrhage and only partially rebounded after VR. The mean Paco2 gradually increased throughout the observation period and was significantly greater than baseline after VR. Mean respiratory rate remained unchanged between baseline and immediately after controlled hemorrhage but was significantly increased from baseline after VR.
Mean ± SD physiologic variables for 7 healthy adult purpose-bred English Pointers immediately before (baseline) and after controlled hemorrhage and immediately after VR with autologous blood.
| Variable | Baseline | Immediately after controlled hemorrhage | Immediately after VR |
|---|---|---|---|
| HR (beats/min) | 110 ± 8a | 103 ± 7b | 118 ± 8c |
| SI (mL/beat/m2) | 34 ± 6a | 20 ± 3b | 36 ± 6a |
| CI (L/min/m2) | 3.71 ± 0.53a | 2.07 ± 0.26b | 4.25 ± 0.63a |
| CVP (mm Hg) | 6 ± 2a | 1 ± 1b | 8 ± 3c |
| PAOP (mm Hg) | 8 ± 3a | 3 ± 2b | 11 ± 4c |
| MAP (mm Hg) | 66 ± 2a | 44 ± 1b | 60 ± 5c |
| SVRI (dynes/s/cm−5·m2) | 1,317 ± 194a | 1,686 ± 208b | 997 ± 196c |
| PI (%) | 1.1 ± 0.7 | 0.6 ± 0.3 | 1.1 ± 0.7 |
| Paco2 (mm Hg) | 46.7 ± 4.5a | 48.4 ± 6.4ab | 49.9 ± 6.0b |
| Pa02 (mm Hg) | 480.7 ± 16.6 | 502.1 ± 33.5 | 485.6 ± 12.1 |
| Hct (%) | 42 ± 5 | 42 ± 4 | 43 ± 5 |
| Total plasma protein (g/dL) | 5.1 ± 0.3a | 4.6 ± 0.4b | 4.7 ± 0.5a,b |
| Arterial blood pH | 7.28 ± 0.03a | 7.24 ± 0.06b | 7.27 ± 0.04a,b |
| Base excess in extracellular fluid (mmol/L) | –5.5 ± 1.6 | –7.0 ± 2.0 | –5.5 ± 1.5 |
| Respiratory rate (breaths/min) | 20 ± 2a | 20 ± 3a | 23 ± 3b |
| Peak inspiratory pressure (cm H2O) | 10 ± 1 | 10 ± 1 | 11 ± 1 |
| Dynamic compliance (mL/cm H2O) | 34 ± 2 | 37 ± 5 | 34 ± 6 |
| Core body temperature (°C) | 37.8 ± 0.5 | 37.8 ± 0.4 | 37.5 ± 0.3 |
Values with different superscript letters differ significantly (P < 0.05).
Each dog was anesthetized with isoflurane and mechanically ventilated following induction of a neuromuscular blockade with atracurium (0.2 mg/kg, IV bolus followed by 0.2 mg/kg/h as a constant rate IV infusion). Prior to controlled hemorrhage, the ETISO was adjusted so that the MAP for each dog ranged between 60 and 70 mm Hg, During controlled hemorrhage, the MAP for each dog was decreased to 40 to 50 mm Hg by adjusting the volume of blood removed. For each dog, the autologous blood that was removed during controlled hemorrhage was retransfused during VR.
PPV and PVI
The mean PI did not differ significantly among collection times (Table 1). The mean ± SD PPV immediately after controlled hemorrhage (20 ± 6%) was significantly greater than that at baseline (9 ± 2%) and immediately after VR (7 ± 3%); however, the mean PPV after VR did not differ significantly from that at baseline (Figure 1). The mean PVI immediately after controlled hemorrhage (18 ± 4%) was significantly greater, whereas that immediately after VR (10 ± 3%) was significantly lower, compared with the mean PVI at baseline (13 ± 2%).
Scatterplots of PPV (A) and PVI (B) for 7 healthy adult purpose-bred English Pointers immediately before (baseline; BL) and after controlled hemorrhage (HV) and immediately after VR with autologous blood. Each dog was anesthetized with isoflurane and mechanically ventilated following induction of a neuromuscular blockade with atracurium (0.2 mg/kg, IV bolus followed by 0.2 mg/kg/h as a constant rate IV infusion). Prior to controlled hemorrhage, the ETISO was adjusted so that the MAP for each dog ranged between 60 and 70 mm Hg. During controlled hemorrhage, the MAP for each dog was decreased to 40 to 50 mm Hg by adjusting the volume of blood removed during controlled hemorrhage. For each dog, the autologous blood that was removed during controlled hemorrhage was retransfused during VR. Each geometric shape represents 1 dog. Within each observation period, the middle horizontal line represents the mean and the brackets represent the SD. a–cValues with different superscript letters differ significantly (P < 0.05).
Citation: American Journal of Veterinary Research 77, 3; 10.2460/ajvr.77.3.280
Scatterplots of PPV (A) and PVI (B) for 7 healthy adult purpose-bred English Pointers immediately before (baseline; BL) and after controlled hemorrhage (HV) and immediately after VR with autologous blood. Each dog was anesthetized with isoflurane and mechanically ventilated following induction of a neuromuscular blockade with atracurium (0.2 mg/kg, IV bolus followed by 0.2 mg/kg/h as a constant rate IV infusion). Prior to controlled hemorrhage, the ETISO was adjusted so that the MAP for each dog ranged between 60 and 70 mm Hg. During controlled hemorrhage, the MAP for each dog was decreased to 40 to 50 mm Hg by adjusting the volume of blood removed during controlled hemorrhage. For each dog, the autologous blood that was removed during controlled hemorrhage was retransfused during VR. Each geometric shape represents 1 dog. Within each observation period, the middle horizontal line represents the mean and the brackets represent the SD. a–cValues with different superscript letters differ significantly (P < 0.05).
Citation: American Journal of Veterinary Research 77, 3; 10.2460/ajvr.77.3.280
Scatterplots of PPV (A) and PVI (B) for 7 healthy adult purpose-bred English Pointers immediately before (baseline; BL) and after controlled hemorrhage (HV) and immediately after VR with autologous blood. Each dog was anesthetized with isoflurane and mechanically ventilated following induction of a neuromuscular blockade with atracurium (0.2 mg/kg, IV bolus followed by 0.2 mg/kg/h as a constant rate IV infusion). Prior to controlled hemorrhage, the ETISO was adjusted so that the MAP for each dog ranged between 60 and 70 mm Hg. During controlled hemorrhage, the MAP for each dog was decreased to 40 to 50 mm Hg by adjusting the volume of blood removed during controlled hemorrhage. For each dog, the autologous blood that was removed during controlled hemorrhage was retransfused during VR. Each geometric shape represents 1 dog. Within each observation period, the middle horizontal line represents the mean and the brackets represent the SD. a–cValues with different superscript letters differ significantly (P < 0.05).
Citation: American Journal of Veterinary Research 77, 3; 10.2460/ajvr.77.3.280
Immediately after controlled hemorrhage, PPV was positively correlated (r2 = 0.77; P < 0.001) with PVI (Figure 2). However, PPV was not correlated with PVI at baseline (r2 = 0.03; P = 0.68) or immediately after VR (r2 = 0.13; P = 0.42).
Scatterplot of PVI versus PPV immediately after controlled hemorrhage for the dogs of Figure 1. The linear regression line is provided, and has an equation of y = 0.52x + 7.24, which is associated with a significant (P < 0.001) positive correlation (r2 = 0.77). See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 77, 3; 10.2460/ajvr.77.3.280
Scatterplot of PVI versus PPV immediately after controlled hemorrhage for the dogs of Figure 1. The linear regression line is provided, and has an equation of y = 0.52x + 7.24, which is associated with a significant (P < 0.001) positive correlation (r2 = 0.77). See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 77, 3; 10.2460/ajvr.77.3.280
Scatterplot of PVI versus PPV immediately after controlled hemorrhage for the dogs of Figure 1. The linear regression line is provided, and has an equation of y = 0.52x + 7.24, which is associated with a significant (P < 0.001) positive correlation (r2 = 0.77). See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 77, 3; 10.2460/ajvr.77.3.280
Discussion
Results of the present study indicated that controlled hemorrhage sufficient to cause hypotension (MAP, 40 to 50 mm Hg) caused a significant increase in PPV and PVI in isoflurane-anesthetized dogs. Increases in PPV induced by hypotension-inducing hemorrhage were consistent with the results of a previous study11 conducted by our laboratory group in which only dogs that became hypotensive (MAP < 60 mm Hg) during controlled hemorrhage (withdrawal of 30% of the estimated blood volume) had a significant increase in PPV Pulse pressure variation and PVI indicated that most dogs of the present study were in the preload-dependent limb of the Frank-Starling curve because VR with autologous blood significantly decreased mean PPV and PVI values from their corresponding values recorded immediately after controlled hemorrhage, whereas mean SI and CI were significantly increased by VR. The explanation for the decrease in PPV and PVI from values recorded after hemorrhage is a minimization of stroke volume oscillations caused by mechanical ventilation subsequent to improvement in cardiac preload induced by VR.
In the present study, mean CVP, PAOP, CI, and SI were significantly decreased from baseline (immediately before controlled hemorrhage) following controlled hemorrhage; however, the mean SI and CI returned to baseline values, whereas the mean CVP and PAOP were significantly increased from baseline immediately after VR. Although those changes were consistent with the concept that CVP and PAOP are indicators of fluid responsiveness, CVP and PAOP are often unreliable for predicting whether CI and SI will increase significantly following volume loading in healthy and septic individuals.13–16 Static preload indices (CVP and PAOP) may have poor correlation with changes in cardiac output because end-diastolic volume (and consequently stroke volume) is determined by intracavitary pressures as well as other factors such as compliance of the cardiac chambers, transmural pressures (ie, the difference between pressures inside and outside the cardiac chambers), and tonus of the venous capacitance vessels.13–17
Heart rate is expected to increase during hypotension-inducing hemorrhage because there is an increase is sympathetic outflow associated with decreased baroreceptor activity. In the present study, the mean HR immediately after controlled hypotension-inducing hemorrhage was significantly decreased from that at baseline; however, the magnitude of that decrease (6%) was not clinically relevant. Isoflurane causes dose-dependent inhibition of the baroreceptor reflex and sympathetic activity.18 For the dogs of the present study, anesthesia was maintained with isoflurane at concentrations (mean ± SD ETISO, 1.74 ± 0.48%) that typically produce a moderate depth of anesthesia when the inhalant anesthetic is not combined with anesthetic adjuvants, and those concentrations were sufficient to suppress the baroreceptor reflex and prevent the compensatory increase in HR that was expected during hypotension-inducing hemorrhage.
Increases in ETISO are associated with decreases in MAP, which are attributed to a decrease in central sympathetic outflow and vasodilation caused by isoflurane.19 In the present study, we chose to titrate the ETISO to maintain MAP between 60 and 70 mm Hg prior to initiation of controlled hemorrhage because use of a fixed ETISO could have resulted in greater variation in cardiovascular variables owing to differences in the sensitivity of individual dogs to the cardiovascular effects of isoflurane, and isoflurane-induced hypotension (MAP < 60 mm Hg) prior to controlled hemorrhage would prevent us from determining the effects of hypotension-inducing hemorrhage on PPV.
In humans, PPV and PVI are not indicators of a patient's volume status, but they are useful for predicting which individuals are likely to respond to fluid therapy with an increase in SI and CI (ie, are in the ascending limb of the Frank-Starling curve regardless of whether they have volume deficits).1–7 Three of the 7 dogs of the present study had a PPV > 15% prior to collection of baseline data. Those dogs were administered tetrastarch (5 mL/kg, IV) to increase circulating blood volume (ie, cardiac preload) and decrease PPV to < 15% so that we could test the hypothesis that hypotension-inducing hemorrhage causes an increase in PPV. Because a PPV threshold to discriminate between dogs that will and will not respond to fluid therapy with increases in SI and CI has not been established, we arbitrarily chose to maintain PPV < 15% prior to initiation of controlled hemorrhage on the basis of results of a study20 that involved pigs in which a PPV > 12% to 15% was predictive of fluid responsiveness.
Although tetrastarch administration to 3 of the 7 dogs of the present study might have increased the circulating blood volume in those dogs, it did not bias our results because controlled hemorrhage was continued until a target MAP (40 to 50 mm Hg) was achieved, and any increase in the circulating blood volume was removed during controlled hemorrhage. Of the 3 dogs that received tetrastarch, the blood volume (35 mL/kg) removed during controlled hemorrhage from 1 was substantially greater than, whereas that (24 and 25 mL/kg) for the other 2 dogs was similar to, the blood volume removed from the 4 dogs that did not receive tetrastarch (range, 19 to 26 mL/kg).
In humans, the correlation between PPV and PVI ranges between 0.50 and 0.72.3,4,6 For the dogs of the present study, there was an overall positive correlation (r2 = 0.77) between PPV and PVI after hypotension-inducing hemorrhage. Albeit statistically significant, changes in mean values recorded between baseline and hypotension-inducing hemorrhage were of smaller magnitude for PVI (from 13 ± 3% at baseline to 18 ± 4% after controlled hemorrhage) than for PPV (from 9 ± 2% at baseline to 20 ± 6% after controlled hemorrhage), which suggested that, in dogs, PVI might be less sensitive than PPV for detection of hypotension-inducing hemorrhage because small oscillations in PVI might be attributed to random effects or noise. The sensitivity and specificity of PPV for discrimination between individuals that are and are not likely to respond to fluid therapy with substantial increases in SI or CI are similar to those of PVI in critically ill human patients who are mechanically ventilated.6 In veterinary medicine, additional studies are necessary to evaluate the sensitivity and specificity of and to determine a threshold value for both PPV and PVI for identification of dogs that are fluid responsive during anesthesia.
For the dogs of the present study, changes in PI were not associated with controlled hemorrhage or VR, and throughout the duration of the observation period, the PI for all dogs remained above 0.2, the minimum value recommended for avoidance of a poor plethysmographic signal and for reliable measurement of PVI.4 In human patients, placement of the pulse oximeter probe on an index finger generally results in satisfactory PI readings, but in many patients, administration of vasopressor drugs such as norepinephrine will cause a decrease in peripheral blood flow, which alters the plethysmographic waveform such that the PI decreases or becomes unmeasurable.21 To our knowledge, no studies have been performed to determine the optimal site for placement of a pulse oximeter probe in dogs for optimization of the PI signal; therefore, in the dogs of the present study, we chose to place a clip-type pulse oximeter probe on the mid portion of the tongue because that is standard practice in small animal practice. On the basis of our clinical experience, prolonged application of pressure by the pulse oximeter clip at 1 site of the tongue can reduce the amplitude of the plethysmographic waveform and cause artificially low oxygen saturation readings because of abnormally decreased perfusion and venous stasis. To avoid those artifacts and optimize the PI signal, we gently massaged the tongue and repositioned the pulse oximeter probe 3 minutes before each data collection.
Dogs were used in a study22 conducted to evaluate changes in PPV during graded hemorrhage as well as in the studies8,9 that pioneered the use of parameters derived from analysis of arterial blood pressure (ie, systolic pressure variation) tracings to estimate changes in cardiac preload during hemorrhage and VR. In those studies,8,22 an inflatable vest was placed around the thorax of each dog to mimic the chest wall-to-lung compliance ratio of humans. Because use of the inflatable vests affected venous return and altered the PPV and systolic pressure variation of the dogs, the results of those studies21–23 have limited clinical applicability to veterinary practice and cannot be directly compared with the results of the present study.
In the present study, tidal volume was held constant (12 mL/kg) throughout the observation period with the dogs under an atracurium-induced neuromuscular blockade to provide additional muscle relaxation and prevent patient-ventilator dyssynchrony. Changes in tidal volume and instances in which the patient has respiratory efforts that are not in synchrony with the mandatory tidal breaths regulated by the ventilator can alter intrapleural pressure and affect venous return, resulting in unreliable calculation of PPV and PVI.23,24
In the present study, limiting the respiratory rate (and consequently the minute volume ventilation) to a maximum of 26 breaths/min resulted in mild hypercapnia (Paco2, 45 to 50 mm Hg) in most dogs, although 1 dog developed moderate hypercapnia (Paco2, 53 to 63 mm Hg). Because the dogs were mechanically ventilated and we did not increase minute volume ventilation to normalize Paco2, the dogs developed acidemia (pH, 7.18 and 7.35) throughout the observation period. A mild increase in Paco2 from 40 to 52 mm Hg may cause a concurrent increase in CI and HR and a decrease in SVRI.25 It is likely that hypercapnia caused some degree of cardiovascular stimulation and altered the cardiovascular variables in the dogs of the present study; however, controlled hemorrhage until a target MAP (40 to 50 mm Hg) was achieved resulted in hemodynamic responses (decrease in CI and increase in SVRI) that were expected secondary to hypovolemia. We chose to limit the respiratory rate to 26 breaths/min instead of preventing hypercapnia because excessive tachypnea might have decreased the expiratory time to such an extent that the PPmin and PImin could not be recorded, which invariably occurs during expiration in mechanically ventilated individuals.
The primary limitation of the present study was the small number of dogs evaluated. One of the 7 dogs failed to develop substantial changes in PPV and PVI immediately after hypotension-inducing hemorrhage or VR, most likely because of factors that were not controlled in the present study. In that dog, the PPV was 7% at baseline, 9% immediately after controlled hemorrhage, and 6% immediately after VR; and the PVI was 14% at baseline and immediately after controlled hemorrhage and 6% immediately after VR. Pulse pressure variation and PVI are affected by changes in the circulating blood volume, arterial compliance, HR and rhythm, tidal volume, chest wall and pulmonary compliance, and right ventricular failure.1 Examination of the cardiovascular and pulmonary mechanics data for that dog did not reveal anything that could explain the absence of substantial changes in PPV and PVI during the observation period. It is possible that the volume of blood removed from that dog during controlled hemorrhage (19 mL/kg; 24% of the estimated blood volume) was insufficient to decrease preload and increase PPV and PVI. A substantial decrease in circulating blood volume can cause venoconstriction, which in turn results in blood centralization, or the shifting of blood flow from the peripheral to central circulation.26,27 However, for the dog with the fairly stable PPV and PVI throughout the duration of the present study, the CVP and PAOP were 4 and 6 mm Hg, respectively, at baseline and decreased to 1 and 2 mm Hg, respectively, immediately after controlled hemorrhage, which was inconsistent with centralization of blood volume.
Findings of the present study indicated that controlled hemorrhage sufficient to cause hypotension (MAP < 50 mm Hg) will cause a significant increase in the PPV and PVI of isoflurane-anesthetized dogs that are mechanically ventilated. The subsequent decrease of PPV and PVI to prehemorrhage values and the normalization of SI and CI after VR suggested that those indices may be useful for identification of dogs that will respond favorably to fluid therapy (ie, dogs in the preload-dependent portion of the Frank-Starling curve). Future studies are necessary to establish the threshold values and to compare the sensitivity and specificity of PPV and PVI for identifying dogs that are responsive to fluid administration during anesthesia.
Acknowledgments
Supported by the São Paulo Research Foundation (FAPESP grant No. 2012/03207-2) and the Brazilian National Council for Scientific and Technological Development (CNPq grant No. 309529/2011-5).
This manuscript represents a portion of a thesis submitted by Dr. Klein to the Department of Anesthesiology, Faculdade de Medicina, Univ. Estadual Paulista (UNESP), as partial fulfillment of the requirements for a Master of Science degree.
ABBREVIATIONS
| CI | Cardiac index |
| CVP | Central venous pressure |
| ETISO | End-tidal concentration of isoflurane |
| HR | Heart rate |
| MAP | Mean arterial pressure |
| PAOP | Pulmonary artery occlusion pressure |
| PI | Perfusion index |
| PImax | Maximum perfusion index |
| PImin | Minimum perfusion index |
| PPmax | Maximum pulse pressure |
| PPmin | Minimum pulse pressure |
| PPV | Pulse pressure variation |
| PVI | Plethysmographic variablity index |
| SI | Stroke index |
| SVRI | Systemic vascular resistance index |
| VR | Volume replacement |
Footnotes
Isoforine, Cristália Prof Quim e Farm LTDA, São Paulo, Brazil.
Dräger Primus, Drägerwerk AG & Co, Lübeck, Germany.
Warmtouch, Mallinkrodt Medical, Pleasanton, Calif.
Infusion Pump LP8x, Digicare Biomedical Technology, Boynton Beach, Fla.
pH/blood-gas analyzer model 348, Siemens Diagnostics, Halstead, Essex, England.
TruWave PX 260, Edwards Lifesciences, Irvine, Calif.
DX 2020 monitor, Dixtal Biomédica–Philips, São Paulo, Brazil.
Invasive blood pressure module DX-ALIBP plus, Dixtal Biomédica–Philips, São Paulo, Brazil.
Ventilation module DX-AJVEN-0, Dixtal Biomédica–Philips, São Paulo, Brazil.
LNCS TC-I, Masimo Corp, Irvine, Calif.
Masimo Radical-7, version 7.7, Masimo Corp, Irvine Calif.
Swan-Ganz Catheter Model 131HF7, Edwards Lifesciences, Irvine, Calif.
AS/3 anesthesia monitor, Datex-Ëngstrom, Helsinki, Finland.
Digipump SR7x, Digicare Biomedical Technology, Boynton Beach, Fla.
Neurostimulator E2107, BGE Médica LTDA, São Paulo, Brazil.
Voluven 6%, Fresenius-Kabi Brasil LTDA, Barueri, Brazil.
CPDA blood collection bag, JP Indústria Farmacêutica SA, Ribeirão Preto, Brazil.
Prism for Windows, version 6.01, GraphPad Software Inc, La Jolla, Calif.
References
1. Michard F. Changes in arterial pressure during mechanical ventilation. Anesthesiology 2005; 103: 419–428.
2. Auler JO Jr, Galas F, Hajjar L, et al. Online monitoring of pulse pressure variation to guide fluid therapy after cardiac surgery. Anesth Analg 2008; 106: 1201–1206.
3. Cannesson M, Delannoy B, Morand A, et al. Does the pleth variability index indicate the respiratory-induced variation in the plethysmogram and arterial pressure waveforms? Anesth Analg 2008; 106: 1189–1194.
4. Chandler JR, Cooke E, Petersen C, et al. Pulse oximeter plethysmograph variation and its relationship to the arterial waveform in mechanically ventilated children. J Clin Monit Comput 2012; 26: 145–151.
5. Marik PE, Cavallazzi R, Vasu T, et al. Dynamic changes in arterial waveform derived variables and fluid responsiveness in mechanically ventilated patients: a systemic review of the literature. Crit Care Med 2000; 37: 2642–2647.
6. Loupec T, Nanadoumgar H, Frasca D, et al. Pleth variability index predicts fluid responsiveness in critically ill patients. Crit Care Med 2011; 39: 294–299.
7. Julien F, Hilly J, Sallah TB, et al. Plethysmographic variability index (PVI) accuracy in predicting fluid responsiveness in anesthetized children. Paediatr Anaesth 2013; 23: 536–546.
8. Perel A, Pizov R, Cotev S. Systolic blood pressure variation is a sensitive indicator of hypovolemia in ventilated dogs subjected to graded hemorrhage. Anesthesiology 1987; 67: 498–502.
9. Szold A, Pizov R, Segal E, et al. The effect of tidal volume and intravascular volume state on systolic pressure variation in ventilated dogs. Intensive Care Med 1989; 15: 368–371.
10. Bennett FM, Tenney SM. Comparative mechanics of mammalian respiratory system. Respir Physiol 1982; 49: 131–140.
11. Diniz MS, Teixeira-Neto FJ, Cândido TD, et al. Effects of dexmedetomidine on pulse pressure variation changes induced by hemorrhage followed by volume replacement with autologous blood in isoflurane anesthetized dogs. J Vet Emerg Crit Care (San Antonio) 2014; 24: 681–692.
12. Shoemaker WC, Parsa MH. Invasive and noninvasive monitoring. In: Shoemaker WC, Ayres SM, Grevnik A, et al, eds. Textbook of critical care. 4th ed. Philadelphia: WB Saunders Co, 2000; 74–91.
13. Michard F, Teboul JL. Predicting fluid responsiveness in ICU patients. Chest 2002; 121: 2000–2008.
14. Kumar A, Anel R, Bunell E, et al. Pulmonary artery occlusion pressure and central venous pressure fail to predict ventricular filling volume, cardiac performance, or the response to volume infusion in normal subjects. Crit Care Med 2004; 32: 691–699.
15. Osman D, Ridel C, Ray P, et al. Cardiac filling pressures are not appropriate to predict hemodynamic response to volume challenge. Crit Care Med 2007; 35: 64–68.
16. Marik PE, Baram M, Vahid B. Does the central venous pressure predict fluid responsiveness? A systematic review of the literature and the tale of seven mares. Chest 2008; 134: 172–178.
17. Marik PE, Monnet X, Teboul JL. Hemodynamic parameters to guide fluid therapy. Ann Intensive Care 2011; 1: 1–9.
18. Muzi M, Ebert TJ. A comparison of baroreflex sensitivity during isoflurane and desflurane anesthesia in humans. Anesthesiology 1995; 82: 919–925.
19. Steffey EP, Howland D Jr. Isoflurane potency in the dog and cat. Am J Vet Res 1977; 38: 1833–1836.
20. Broch O, Gruenewald M, Renner J, et al. Dynamic and volumetric variables reliably predict fluid responsiveness in a porcine model with pleural effusion. PLoS ONE 2013; 8: e56267.
21. Monnet X, Guérin L, Jozwiak M, et al. Pleth variability index is a weak predictor of fluid responsiveness in patients receiving norepinephrine. Br J Anaesth 2013; 110: 207–213.
22. Berkenstadt H, Friedman Z, Preisman S, et al. Pulse pressure and stroke volume variations during severe haemorrhage in ventilated dogs. Br J Anaesth 2005; 94: 721–726.
23. Kim HK, Pinsky MR. Effect of tidal volume, sampling duration, and cardiac contractility on pulse pressure and stroke volume variation during positive-pressure ventilation. Crit Care Med 2008; 36: 2858–2862.
24. Desebbe O, Boucau C, Farhat F, et al. The ability of pleth variability index to predict the hemodynamic effects of positive end-expiratory pressure in mechanically ventilated patients under general anesthesia. Anesth Analg 2010; 110: 792–798.
25. Mas A, Saura P, Joseph D, et al. Effect of acute moderate changes in PaCO2 on global hemodynamics and gastric perfusion. Crit Care Med 2000; 28: 360–365.
26. Nouira S, Elatrous S, Dimassi S, et al. Effects of norepinephrine on static and dynamic preload indicators in experimental hemorrhagic shock. Crit Care Med 2005; 33: 2339–2343.
27. Bouchacourt JP, Riva JA, Grignola JC. The increase of vasomotor tone avoids the ability of the dynamic preload indicators to estimate fluid responsiveness. BMC Anesthesiol 2013; 13: 41.
