Abstract
OBJECTIVE
To evaluate the diagnostic usefulness of focused cardiac ultrasonography and selected echocardiographic variables for predicting fluid responsiveness in conscious, spontaneously breathing dogs with various clinical conditions.
ANIMALS
26 dogs (15 males and 11 females) with a median age of 84 months (range, 12 to 360 months) and median body weight of 8 kg (range, 2 to 35 kg) referred for various clinical conditions.
PROCEDURES
Left ventricular end-diastolic internal diameter normalized to body weight (LVIDDn), left ventricular volume score (LVVS), left ventricular end-diastolic volume index (EDVI), aortic velocity time integral (VTIAo), and aortic peak flow velocity (VmaxAo) were echocardiographically measured before and after IV administration of a bolus of lactated Ringer solution (4 mL/kg) over a 1-minute period. Dogs were classified on the basis of the observed change in aortic stroke volume following fluid administration as responders (≥ 15%) or nonresponders (< 15%) to fluid administration. Receiver operating characteristic curves were generated for the ability of LVVS, LVIDDn, EDVI, VTIAo, and VmaxAo to predict responder status.
RESULTS
13 dogs were classified as responders and 13 as nonresponders. Areas under the receiver operating characteristic curves (95% confidence intervals) for predicting fluid responsiveness were as follows: VTIAo, 0.91 (0.74 to 0.99); LVIDDn, 0.85 (0.66 to 0.96); EDVI, 0.85 (0.65 to 0.96); LVVS, 0.85 (0.65 to 0.96); and VmaxAo, 0.75 (0.54 to 0.90).
CONCLUSIONS AND CLINICAL RELEVANCE
The evaluated echocardiographic variables were useful for noninvasive prediction of fluid responsiveness in conscious dogs and could be valuable for informing clinical decisions regarding fluid therapy.
Intravenous fluid administration to maintain an adequate intravascular volume status and optimize cardiac output is a key aspect of hemodynamic resuscitation in humans1 and veterinary species.2 Severe volume depletion has severe adverse consequences and must be identified and corrected quickly.3 Often hypovolemia is associated with hypotension, which can increase the risk of morbidity and death in surgical and critically ill human patients.4,5 Adequate early fluid resuscitation of human patients may improve the likelihood of survival,6,7 whereas excessive fluid administration increases the risk of death.8–10 Nevertheless, fluid challenge (IV administration of a small amount of fluid over a brief period) as a method to reveal a fluid volume deficit should be considered as having potentially detrimental effects in patients with no fluid needs.
In a study11 involving healthy human volunteers, IV infusion of an amount of lactated Ringer solution equivalent to that used routinely during minor to moderate surgical procedures resulted in a significant decrease in pulmonary function and corresponding weight gain. Fluid therapy guidelines for dogs indicate that a shock dose (80 to 90 mL/kg, IV) of crystalloid fluid should be administered to severely hypovolemic patients, starting by rapidly administering 25% of the calculated dose and monitoring the dog's response (eg, resolution of clinical signs of hypoperfusion such as persistently high heart rate, poor pulse quality, hypotension, and low urine output).12 Fluid therapy must be stopped if the dog develops any signs of fluid overload, such as an increase in respiratory rate and effort, peripheral or pulmonary edema, and weight gain.12 However, such signs may be lagging markers and could appear only in advanced fluid overload conditions, resulting in indicators of organ dysfunction (ie, pulmonary edema).
Clinical examination and routine cardiovascular monitoring are considered inadequate for assessing hemodynamic and intravascular volume status,1 as are traditionally adopted static indices of cardiac preload such as CVP and PAOP.13,14 Therefore, reliable and clinically easy-to-use approaches are needed that are able to predict a patient's fluid needs.
Fluid responsiveness is the ability of an individual to increase their stroke volume by 10% to 15% after receiving an appropriate fluid bolus. For evaluation of FR in humans, dynamic indices are reportedly superior to static volumetric (eg, LVEDA) and pressure (eg, CVP or PAOP) indices.15 However, dynamic indices such as SVV, pulse pressure variation, and SPV perform better than static indices in anesthetized and mechanically ventilated humans with a constant intrathoracic pressure variation between breaths.16
To the authors' knowledge, few indices of FR have been evaluated in small animals. Systolic pressure variation is an index that can purportedly predict a 10% increase in mean arterial blood pressure and decrease in heart rate in healthy, mechanically ventilated, anesthetized dogs.17 This index is also well correlated with the ultrasonographically determined ratios between the diameter of the caudal vena cava and the diameter of the aorta in dogs.18 Another dynamic index, respiratory variation of VmaxAo, has been reported as an optimal predictor of FR (area under the ROC curve, 0.95) in anesthetized and mechanically ventilated dogs.19 However, SPV and respiratory variation of aortic flow peak velocity can be evaluated only in anesthetized dogs that have completely adapted to controlled mechanical ventilation, and in spontaneously breathing subjects, in which breath depth cannot be controlled, dynamic indices are unreliable. Unfortunately, fluid resuscitation is commonly initiated and reassessed in conscious patients in emergency care or intensive care settings.
For these reasons, FoCUD represents a potentially useful technique for assessing FR in canine patients. Echocardiographic measurements of FR can be obtained rapidly, noninvasively, and without pain to subjects and are fairly simple to learn to perform. Moreover, echocardiography can also be used to evaluate changes in SV, correct measurement of which is important when attempting to determine the discriminatory power of an FR index. Indeed, echocardiography can be used to measure the SVV from heartbeat to heartbeat and intrasubject variation in VTIAo.20–22
Some static echocardiographic indices have been evaluated in conscious humans and proposed for use in cardiac preload assessment. Left ventricular end-diastolic volume, LVEDA, and end-systolic cavity obliteration (also referred to as the kissing papillary muscles sign) are useful for identifying hypovolemia.23–25 Both LVEDV and LVEDA are considered static indices because they are not influenced by the heart-lung interaction that induces cyclic variation in cardiac preload and they offer the opportunity to derive indices of FR that can be measured in conscious dogs. To the authors' knowledge, no indices have been identified that accurately predict FR in spontaneously breathing dogs.
The purpose of the study reported here was to assess the discriminatory power of FoCUD and selected echocardiographic indices to predict FR in spontaneously breathing dogs. Specifically, LVIDDn, left ventricular EDVI, VTIAo, and VmaxAo were explored as independent predictors of FR in a heterogeneous group of conscious, critically ill dogs.
Materials and Methods
Animals
Dogs referred to Centro Veterinario Imperiese and Policlinico Veterinario Roma Sud for various clinical conditions from July 2013 to March 2014 were considered for inclusion in this prospective, 2-center study. Dogs were excluded if they were < 1 year of age, had a history of cardiovascular disease or endocrine disorders, or were receiving vasoactive drug treatment. All owners were informed about the study and provided consent to proceed.
Procedures
For each dog, a thorough clinical examination and basic hematologic analysis (Hct and serum total protein and electrolyte concentrations and blood gas analysis) were performed. Heart rate, respiratory rate, and arterial blood pressure were measured via 3 consecutive readings. Arterial blood pressure was measured with a high-definition oscillometric systema on a peripheral artery with the dog positioned in right lateral recumbency and the cuff applied on the right forelimb. An IV catheter was placed in a cephalic vein. Echocardiography was then performed before and after a bolus of lactated Ringer solution was administered IV over a 1-minute period by means of 1 or more preloaded 50-mL syringes through a catheter placed in a cephalic vein.
Echocardiography
Echocardiographic examinations were performed by 1 operator at each center (RR or SO) by use of an ultrasound machineb,c and 2- to 4-MHz, 5- to 7.5-MHz, and 7.5- to 10-MHz phased array probes,d–f with dogs positioned in right lateral recumbency. Images and cine loops were acquired in right parasternal long-axis and short-axis and subcostal views. Throughout the examination, the ECG tracing was recorded. An M-mode evaluation of the left ventricle was performed by placement of the M-mode cursor between the papillary muscles; LVEDD and left ventricular end-systolic diameter were measured with the leading edge–to–leading edge method. A pulsed-wave Doppler evaluation of aortic flow was conducted by placement of the Doppler volume sampler on the aortic valve; the VTIAo and VmaxAo were also measured. The subcostal view was the preferred approach (Figure 1).26
Representative subcostal standard echocardiographic view of the left ventricular outflow tract in a dog that has been optimized for imaging that tract. The beat-to-beat VTIAo before and after the volume expansion was recorded, and the median VTIAo was calculated over 1 respiratory cycle.
Citation: American Journal of Veterinary Research 80, 4; 10.2460/ajvr.80.4.369
The duration of the entire procedure was recorded. All images were archived, and measurements were performed in a later session. To avoid any influence of heart-lung interaction on the results, during the Doppler evaluation, all consecutive heartbeats over 1 respiratory cycle were recorded, and the R-R intervals preceding every left ventricular ejection were measured.22 All M-mode measurements were performed at the end of expiration, and 3 measurements of 3 consecutive respiratory cycles were performed. The median value of each echocardiographic and Doppler measurement was reported.
Echocardiographic indices
All images and cine loops were archived, and measurements of VTIAo and VmaxAo with R-R intervals and LVEDD were performed by 2 observers (SO or RR). The ‡SVAo was calculated as follows: (VTIAo after bolus administration – VTIAo before bolus administration)/VTIAo before bolus administration. Dogs were then classified as responders (‡SVAo ≥ 15%) or nonresponders (‡SVAo < 15%) to bolus IV fluid administration.27 Values for LVEDD were scaled to individual dogs by use of the allometric scaling method proposed by Cornell et al,28 and findings regarding left ventricular preload were used to assign an LVVS as an index of circulating blood volume. Specifically, dogs were assigned an LVVS of 1 (hypervolemia) if they had an LVEDD greater than the Cornell maximum value, an LVVS of 0 (normovolemia) if they had an LVEDD between the Cornell minimum and maximum values, an LVVS of −1 (hypovolemia) if they had an LVEDD less than the Cornell minimum value, and an LVVS of −2 (severe hypovolemia) if they had an LVEDD less than the Cornell minimum value as well as end-systolic cavity obliteration with no left ventricular end-systolic diameter (ie, the left ventricular free wall and the interventricular septum endocardial borders were touching).
The LVIDDn, a recently proposed echocardiographic index of cardiac size,28,29 was calculated for each dog as LVEDD/(body weight)0.294 on the basis of the constant derived from the regression analyses of Cornell et al.28 The LVEDV was calculated as follows: (7 × LVEDD3)/(2.4 + LVEDD).30 The left ventricular EDVI was then calculated as LVEDV/body surface area, whereby body surface area was calculated as 0.101 × (body weight)2/3.
Statistical analysis
Continuous variables were evaluated for normal data distribution by visual inspection of histograms and use of the Shapiro-Wilk test. All values were nonnormally distributed and are therefore reported as median (range). Differences between nonnormally distributed data were analyzed with the Mann-Whitney U test for 2 independent samples and the Wilcoxon signed rank test for 2 related samples. Categorical data are reported as counts and percentages.
Intra- and interobserver agreement in LVEDD measurements was evaluated for 2 operators (SO and RR), blinded to dog identity, who performed these measurements on recorded images for 10 dogs that were not involved in this study and were preliminarily tested. Interobserver reliability was evaluated by calculation of the single-measure ICC with a 2-way mixed-effects model (single measures) to assess absolute agreement in measurements. Intraobserver reliability was measured with a 2-way mixed-effects model (repeated measures) yielding single-measure ICCs for test-retest reliability. The CV was calculated with the logarithmic method. Inter- and intraobserver reliability in VTIAo and VmaxAo measurements was assessed by having the same 2 blinded operators perform duplicated measurements of recorded images for the first 10 dogs in the study and calculating the ICC. An ICC value > 0.7 was used to indicate sufficient reliability.31
To assess the ability of LVVS, EDVI, LVIDDn, VTIAo, and VmaxAo to discriminate between fluid responder and nonresponder status as previously defined, an ROC curve for each predictor variable was generated by use of the nonparametric method. A plot of each observer's true-positive rate (sensitivity) versus the false-positive rate (1 – specificity) was created, and the optimal cutoff value corresponding to the best compromise between the true- and false-positive rates was assessed by calculation of the Youden index value. The area under the ROC curve could have taken on any value between 0 and 1, and the null hypothesis was that it would be 0.5. The range of values for each independent variable for which responder status could not be reliably predicted (sensitivity or specificity < 90%; diagnosis tolerance of 10%) was defined as the gray zone.32 Values within this range would not allow a clinician to determine whether FR was present or absent; therefore, it would be important that this range be identified. An a priori sample size calculation indicated that the minimum number of dogs required to obtain a significant area under the ROC curve with good discriminatory accuracy (> 0.80; α = 0.05 and β = 0.20) was 20. Positive and negative predictive values were calculated on the basis of the FR rate identified in the study. Statistical softwareg was used for all analyses, and values of P < 0.05 were considered significant.
Results
Animals
Twenty-eight dogs with various clinical conditions were enrolled in the study. Two dogs were excluded because of a limited number of images of sufficient diagnostic quality. Therefore, the results from 26 dogs (15 males and 11 females) were analyzed.
Median age of the included dogs was 84 months (range, 12 to 360 months), and median body weight was 8 kg (range, 2 to 35 kg). Sixteen dogs were mixed-breed dogs, 2 were Maltese, and there was 1 each of Beagle, Boxer, Doberman Pinscher, Fox Terrier, Russell Terrier, Jagdterrier, West Highland White Terrier, and Yorkshire Terrier. Clinical conditions included polytrauma (n = 4), gastroenteritis (3), pneumothorax (2), dystocia (2), bite wounds (2), diabetic ketoacidosis (2), poisoning (2), cholangiohepatitis (2), meningitis (1), hemoperitoneum (1), pancreatitis (1), vestibular syndrome (1), pleural effusion (1), anemia (1), and polyneuropathy (1).
Procedural findings
The median duration of the entire experimental procedure (prebolus data collection, bolus administration, and postbolus data collection) was 9 minutes (range, 8 to 12 minutes). The median time required to complete the echocardiographic examination was 8 minutes (range, 7 to 10 minutes). Single-measure (absolute agreement) ICCs for intra- and interobserver agreement and the CV for the variables LVEDD, VmaxAo, and VTIAo were summarized (Table 1).
Intraobserver and interobserver agreement in echocardiographic measurements of VTIAo, LVEDD, and VmaxAo in 10 spontaneously breathing dogs with various clinical conditions.
| Agreement type, by variable | CV (%) | ICC |
|---|---|---|
| VTIAo | ||
| Intraobserver | 4.05 | 0.95 (0.82–0.99) |
| Interobserver | — | 0.94 (0.80–0.98) |
| LVEDD | ||
| Intraobserver | 5.43 | 0.97 (0.90–0.99) |
| Interobserver | — | 0.96 (0.85–0.99) |
| VmaxAo | ||
| Intraobserver | 4.78 | 0.98 (0.91–0.99) |
| Interobserver | — | 0.94 (0.80–0.98) |
Data represent results for 2 observers unaware of dog identity who performed duplicate measurements. Parenthetical values for ICC represent 95% confidence intervals.
Predictors of FR
Thirteen of the 26 (50%) dogs were classified as responders to bolus IV fluid administration, and the other 13 (50%) were classified as nonresponders. Data were acquired for all dogs for all variables except arterial blood pressure, which was not recorded for 3 dogs (1 responder and 2 nonresponders) before bolus administration and 14 dogs (7 responders and 7 nonresponders) after bolus administration. The median (range) change in VTIAo (‡SVAo) after bolus administration was 45.6% (20.0% to 105.9%) for responders (P < 0.001), with no significant (P = 0.17) change in nonresponders (median [range] change, −8.0% [−20.3% to 13.3%]; Figure 2).
Box-and-whisker plots and superimposed dot plots of paired VTIAo measurements (before and after bolus administration) in dogs with various clinical conditions that responded (n = 13) or had no response (13) to IV administration of a bolus of lactated Ringer solution (4 mL/kg) over a 1-minute period. The central horizontal line within each box represents the median value, the top and bottom boundaries of each box represent the third and first interquartile values, and the top and bottom whiskers represent the maximum and minimum values.
Citation: American Journal of Veterinary Research 80, 4; 10.2460/ajvr.80.4.369
Hemodynamic and echocardiographic data for repeated measures (before and after bolus administration) in the responder and nonresponder groups were summarized (Table 2). Respective median numbers of consecutive heartbeats analyzed before and after bolus administration were 5 (range, 4 to 9) and 5 (range, 3 to 7) for responders and 4 (range, 3 to 5) and 4 (range, 3 to 7) for nonresponders.
Median (range) hemodynamic and echocardiographic values for dogs with various clinical conditions that responded (n = 13) or had no response (13) to IV administration of a bolus of lactated Ringer solution (4 mL/kg) over a 1-minute period.
| Before bolus | After bolus | |||||
|---|---|---|---|---|---|---|
| Variable | Responders | Nonresponders | P value* | Responders | Nonresponders | P value* |
| HR (beats/min) | 148 (72 to 191) | 74 (59 to 161) | 0.002 | 137 (81 to 182) | 90 (55 to 216) | 0.02 |
| MAP (mm Hg) | 98 (70 to 129) | 105 (70 to 125) | 0.07 | 91 (75 to 113) | 110 (82 to 145) | 0.26 |
| LVVS | −1 (−1 to 0) | 0 (0 to 1) | — | 0 (−1 to 0) | 0 (0 to 1) | — |
| LVIDDn (cm/kg0.294) | 1.08 (0.67 to 1.69) | 1.53 (1.48 to 1.89) | 0.01 | 1.30 (0.77 to 1.77) | 1.63 (1.31 to 1.94) | 0.002 |
| EDVI (mL/m2) | 33 (9 to 93) | 76 (61 to 124) | 0.003 | 50 (15 to 105) | 83 (55 to 128) | 0.004 |
| VTIAo (cm) | 6.5 (3.4 to 12.7) | 12.1 (10.4 to 18.5) | < 0.001 | 9.9 (6.0 to 18.0) | 12.1 (8.3 to 17.0) | 0.22 |
| VmaxAo (m/s) | 0.81 (0.65 to 1.55) | 1.08 (0.79 to 1.82) | 0.03 | 0.97 (0.76 to 1.89) | 1.11 (0.75 to 1.69) | 0.54 |
P values represent results of the Mann-Whitney U test for comparisons between responders and nonresponders. — = Not calculated. HR = Heart rate. MAP = Mean arterial blood pressure.
After bolus administration, heart rate in nonresponders increased from a median of 74 beats/min (range, 59 to 161 beats/min) to 90 beats/min (range, 55 to 216 beats/min), but this difference was not significant (P = 0.22). In responders, heart rate decreased from a median of 148 beats/min (range, 72 to 191 beats/min) to 137 beats/min (range, 81 to 182 beats/min), and this difference was also not significant (P = 0.25).
Areas under the ROC curve for the evaluated predictor variables ranged from 0.75 (VmaxAo; P = 0.01) to 0.91 (VTIAo; P < 0.001; Figure 3). Optimal cutoffs (ie, those that yielded the highest specificity and sensitivity for predicting FR as evaluated by ROC analysis) for continuous predictor variables were ≤ 10.30 cm for VTIAo (specificity and sensitivity, 100% and 84.6%), ≤ 1.34 cm/kg0.294 for LVIDDn (100% and 76.9%), ≤ 55 mL/m2 for EDVI (100% and 76.9%), ≤ −1 for LVVS (100% and 69.3%), and ≤ 0.88 m/s for VmaxAo (92.3% and 61.5%).
Receiver operating characteristic curves comparing the discriminatory ability of various hemodynamic variables to predict FR in the dogs of Figure 2. Values as measured by 2 observers are represented. AUC = Area under the ROC curve. CI = Confidence interval.
Citation: American Journal of Veterinary Research 80, 4; 10.2460/ajvr.80.4.369
The gray zone intervals for continuous predictor variables were 10.45 to 11.86 cm for VTIAo (Figure 4), 1.48 to 1.67 cm/kg0.294 for LVIDDn, 63.1 to 90.1 mL/m2 for EDVI, and 0.86 to 1.34 m/s for VmaxAo. Positive predictive values for all predictor variables were 100%, except for VmaxAo, which had a value of 88.9%. Negative predictive values for VTIAo, LVIDDn, EDVI, LVVS, and VmaxAo were 86.5%, 81.2%, 81.2%, 76.4%, and 70.6%, respectively.
Plot of sensitivity (blue line) and specificity (green line) of VTIAo at various cutoffs for discriminating between responder and nonresponder status for the dogs of Figure 2. The interval of values between a sensitivity or specificity < 90% was defined as the gray zone (gray-shaded region).
Citation: American Journal of Veterinary Research 80, 4; 10.2460/ajvr.80.4.369
Discussion
All echographic indices evaluated in the present study were good to excellent predictors of FR (as defined) in a heterogeneous group of conscious dogs with various clinical conditions. The optimal cutoffs for discriminating responders to bolus fluid administration from nonresponders had 100% specificity for EDVI, LVIDDn, LVVS, and VTIAo. This finding suggested the high accuracy of the test in terms of specificity and the ability to classify dogs as responders without false-positive results. If one of these indices classified a dog as preload dependent (ie, a fluid responder) with very high probability, the dog could be expected to respond with an increase in stroke volume after volume expansion through IV fluid administration. We believe that our data are clinically relevant considering the challenge posed by fluid management in conscious, ill subjects.11
Although abundant evidence indicates that indices with an established cutoff for predicting FR could facilitate provision of individualized fluid therapy, such indices represent a simplification of much more complex systems. Few physiologic systems are binary in nature and the cardiovascular response (which follows the Frank-Staling law) to bolus IV fluid administration is not that of a binary system. Use of a cutoff simplifies the interpretation of complex physiologic events by allowing the creation of logistic mathematical models to predict FR, which perform well for subjects with an all-or-none type of response to IV fluid administration. Unfortunately, many subjects have a less binary response, with values of indices predicting FR close to determined cutoffs. Values around the cutoffs typically are considered to fall within the gray zone, and it is difficult to predict whether subjects with values in this gray zone would be responders or nonresponders. All studies regarding FR, including the present study, have this modeling-related limitation.
We suggest, as others have for humans,32 that gray-zone values for the echocardiographic variables evaluated in our study be used when making clinical decisions about potential fluid responsiveness, allowing greater confidence in values that fall outside this zone, but a more cautionary approach for dogs with values that fall within this zone (Figure 4). Notably, a normovolemic dog with a normally performing myocardium might be predicted to behave as a responder if it receives an adequate fluid bolus; however, FR and hypovolemia are not synonymous. Indeed, a fluid responder is not necessarily hypovolemic and the decision to provide or not provide fluids in a fluid-responsive patient should always be based on the patient's clinical condition.
The variables CVP and PAOP are not good predictors of FR in humans,13,14 nor are they in dogs, as shown in a study33 in which an increase in cardiac output was observed in both normotensive and hypotensive dogs following administration of a 3-mL/kg bolus of lactated Ringer solution. This may be explained by the fact that CVP, PAOP, and arterial blood pressure, defined as static indices, are measures of pressure; their relationship with blood volume is influenced by compliance of the heart and blood vessels and by cardiac output, and therefore, they do not correctly reflect cardiac preload. Dynamic indices such as pulse pressure variation, SPV, and SVV have been proposed as better choices for guiding fluid therapy and have been evaluated in healthy anesthetized, mechanically ventilated dogs.17,34–36 Dynamic indices are more reliable when mechanical ventilation produces a constant and cyclic preload challenge, as achieved in a completely adapted anesthetized patient. These indices are less reliable when used for conscious, spontaneously breathing subjects because they are dependent on tidal volume.37
All echocardiographic indices evaluated in the study reported here were static in nature but performed well for predicting FR in a group of critically ill dogs. Critically ill dogs with hemodynamic instability may have clinical signs such as tachycardia, weak pulse, high capillary refill time, and pale mucous membranes, all of which are clinical variables that might or might not be indicative of a reduction in cardiac preload because of the reported low positive predictive value of these indices in preload status assessment. Fluid administration for this type of patient is simultaneously difficult and crucial. Values of the indices evaluated in the present study can, in only minutes, suggest to clinicians the probability that a fluid bolus will be beneficial for a spontaneously breathing dog. In the present study, half of the included dogs responded to bolus IV fluid administration, with a median increase in stroke volume close to 50%. Dogs in the nonresponder group had no significant change in stroke volume.
The VTIAo is an important index of stroke volume. Multiplying VTIAo by the cross-sectional area of the aortic annulus, obtained by measuring its diameter in early systole (in the parasternal long-axis view), is one of the most common methods used to calculate cardiac output via echocardiography.38 Because the size of a subject's aortic root does not change, aortic blood flow variation is directly correlated with the change in stroke volume. Previous studies38,39 have shown a high correlation between VTIAo variation, as measured via transthoracic echocardiography, and SVV in humans. On the basis of the data for dogs in the present study, the median value of prebolus VTIAo measured over all consecutive beats during 1 respiratory cycle was an excellent index of FR regardless of dog body weight.
Although no detrimental hemodynamic effects due to fluid administration were noted in the present study, 4 dogs in the nonresponder group had a postbolus EDVI > 100, which was a clear sign of ventricular overload. The EDVI is an echocardiographic index used to measure left ventricular volume at the end of diastole scaled to body surface area. End-diastolic volume can be calculated with the Teichholz method, Simpson derived method of disks, and length-area method.40 The Teichholz method involves use of the LVEDD measured by M-mode echocardiography,30 but this method generally overestimates the EDVI, which is more pronounced in subjects with cardiac remodeling41 and requires geometric assumptions that cannot be completely satisfied in such subjects.40 The EDVI does, however, retain some advantages as an index of FR: it is simple to obtain, can be quickly measured, is well known among many clinicians, and has good reliability for test-retest patient assessment.
Similar to the EDVI, the LVVS and LVIDDn are derived from the LVEDD as echocardiographically measured, with the assumption that LVVS and LVIDDn should fall within the range of values of the allometric scaling system proposed by Cornell et al.28 By the LVVS system, responders to fluid administration are scored as −1 or −2. Responders scored as −2 are those with the most severe volume depletion, which may require prompt treatment. This score also allows for prompt recognition of dogs in this condition, because the kissing papillary muscles sign of end-systolic cavity obliteration can be rapidly and easily spotted by clinicians without the need for measurements.42 By simple pattern recognition, clinicians can use FoCUD to quickly gather information that is essential for time-sensitive clinical decision-making. However, such an approach should be used only to answer focused clinical questions, which must be interpreted in light of the given clinical scenario.43
Dogs with an LVVS score of 1, an LVIDDn > 1.67 cm/kg0.294, or an EDVI > 90.1 mL/m2 in the present study were likely hypervolemic; in such dogs, IV fluid administration could be associated with a higher probability of detrimental effects.9,44,45 Clinicians should be aware that hypervolemic dogs may or may not be fluid tolerant. After bolus fluid administration, dogs that have a further increase in stroke volume without a relevant increase in ventricular filling pressure and extravascular lung water can be defined as fluid-tolerant patients. In contrast, dogs that are unable to increase and sometimes substantially decrease their stroke volume, observed as a rise in ventricular filling pressure and extravascular lung water, can be defined as fluid intolerant.46 These dogs, when affected by more severe clinical conditions such as sepsis and endothelial damage, appear more likely than others to develop negative sequela from fluid overload.9,44,45 In humans, lung ultrasonography has been proposed for monitoring for early signs of fluid overload.43
A limitation of the present study was the small sample size and missing data for some dogs, which limited our ability to investigate changes in hemodynamic variables such as heart rate and arterial blood pressure following fluid administration. In addition, we evaluated only static indices of FR, by which the response observed (change in SV) was mainly based on the response of the left ventricle to the administered fluid bolus. Therefore, the influence of other factors involving right ventricular function on FR could not be excluded. We must stress the concept that FR is always a biventricular relationship, and it would be interesting to investigate FR in dogs with a flatter Frank-Starling curve than that of cardiac patients. In addition, it remains unknown whether the echocardiographic methods used would indeed be simple for clinicians with limited or no ultrasonography experience to learn. Despite these limitations, the echocardiographic indices evaluated in the present study appeared to be valuable and feasible for noninvasive prediction of FR in conscious dogs with various clinical conditions.
Acknowledgments
This study was not supported by any grant.
ABBREVIATIONS
| CV | Coefficient of variation |
| CVP | Central venous pressure |
| ‡SVAo | Variation of aortic stroke volume |
| EDVI | End-diastolic volume index |
| FoCUD | Focused cardiac ultrasonography in dogs |
| FR | Fluid responsiveness |
| ICC | Intraclass correlation coefficient |
| LVEDA | Left ventricular end-diastolic area |
| LVEDD | Left ventricular end-diastolic diameter |
| LVEDV | Left ventricular end-diastolic volume |
| LVIDDn | Left ventricular internal diameter in diastole normalized to body weight |
| LVVS | Left ventricular volume score |
| PAOP | Pulmonary artery occlusion pressure |
| ROC | Receiver operating characteristic |
| SPV | Systolic pressure variation |
| SVV | Stroke volume variation |
| VmaxAo | Aortic peak flow velocity |
| VTIAo | Aortic velocity time integral |
Footnotes
VET HDO monitor, S + B medVET GmbH, Babenhausen, Germany.
MyLab class C, Esaote SpA, Genova, Italy.
MyLab 70 CV, Esaote SpA, Genova, Italy.
PA240 probe, Esaote SpA, Genova, Italy.
PA122 probe, Esaote SpA, Genova, Italy.
PA023 probe, Esaote SpA, Genova, Italy.
MedCalc Statistical Software, version 17.9.6, MedCalc Software bvba, Ostend, Belgium.
References
1. Vincent JL, Weil MH. Fluid challenge revisited. Crit Care Med 2006;34:1333–1337.
2. Mandell DC, King LG. Fluid therapy in shock. Vet Clin North Am Small Anim Pract 1998;28:623–644.
3. Byers CG. Fluid therapy: options and rational selection. Vet Clin North Am Small Anim Pract 2017;47:359–371.
4. Shoemaker WC, Wo CC, Thangathurai D, et al. Hemodynamic patterns of survivors and nonsurvivors during high risk elective surgical operations. World J Surg 1999;23:1264–1270.
5. Zenati MS, Billiar TR, Townsend RN, et al. A brief episode of hypotension increases mortality in critically ill trauma patients. J Trauma 2002;53:232–236.
6. Rivers E, Nguyen B, Havstad S, et al. Early goal-directed therapy in the management of severe sepsis and septic shock. N Engl J Med 2001;345:1368–1377.
7. Han YY, Carcillo JA, Dragotta MA, et al. Early reversal of pediatric-neonatal septic shock by community physicians is associated with improved outcome. Pediatrics 2003;112:793–799.
8. Rosenberg AL, Dechert RE, Park PK, et al. Review of a large clinical series: association of cumulative fluid balance on outcome in acute lung injury: a retrospective review of the ARDSnet tidal volume study cohort. J Intensive Care Med 2009;24:35–46.
9. Murphy CV, Schramm GE, Doherty JA, et al. The importance of fluid management in acute lung injury secondary to septic shock. Chest 2009;136:102–109.
10. Boyd JH, Forbes J, Nakada TA, et al. Fluid resuscitation in septic shock: a positive fluid balance and elevated central venous pressure are associated with increased mortality. Crit Care Med 2011;39:259–265.
11. Holte K, Jensen P, Kehlet H. Physiologic effects of intravenous fluid administration in healthy volunteers. Anesth Analg 2003;96:1504–1509.
12. Davis H, Jensen T, Johnson A, et al. 2013 AAHA/AAFP fluid therapy guidelines for dogs and cats. J Am Anim Hosp Assoc 2013;49:149–159.
13. Michard F, Teboul JL. Predicting fluid responsiveness in ICU patients: a critical analysis of the evidence. Chest 2002;121:2000–2008.
14. Marik PE, Baram M, Vahid B. Does central venous pressure predict fluid responsiveness? A systematic review of the literature and the tale of seven mares. Chest 2008;134:172–178.
15. Preisman S, Kogan S, Berkenstadt H, et al. Predicting fluid responsiveness in patients undergoing cardiac surgery: functional haemodynamic parameters including the Respiratory Systolic Variation Test and static preload indicators. Br J Anaesth 2005;95:746–755.
16. Monnet X, Marik PE, Teboul JL. Prediction of fluid responsiveness: an update. Ann Intensive Care 2016;6:111.
17. Rabozzi R, Franci P. Use of systolic pressure variation to predict the cardiovascular response to mini-fluid challenge in anaesthetised dogs. Vet J 2014;202:367–371.
18. Meneghini C, Rabozzi R, Franci P. Correlation of the ratio of caudal vena cava diameter and aorta diameter with systolic pressure variation in anesthetized dogs. Am J Vet Res 2016;77:137–143.
19. Bucci M, Rabozzi R, Guglielmini C, et al. Respiratory variation in aortic blood peak velocity and caudal vena cava diameter can predict fluid responsiveness in anaesthetised and mechanically ventilated dogs. Vet J 2017;227:30–35.
20. Pereira de Souza Neto E, Grousson S, Duflo F, et al. Predicting fluid responsiveness in mechanically ventilated children under general anaesthesia using dynamic parameters and transthoracic echocardiography. Br J Anaesth 2011;106:856–864.
21. Muller L, Toumi M, Bousquet PJ, et al. An increase in aortic blood flow after an infusion of 100 ml colloid over 1 minute can predict fluid responsiveness: the mini-fluid challenge study. Anesthesiology 2011;115:541–547.
22. Slama M, Masson H, Teboul JL, et al. Respiratory variations of aortic VTI: a new index of hypovolemia and fluid responsiveness. Am J Physiol Heart Circ Physiol 2002;283:H1729–H1733.
23. Swenson JD, Harkin C, Pace NL, et al. Transesophageal echocardiography: an objective tool in defining maximum ventricular response to intravenous fluid therapy. Anesth Analg 1996;83:1149–1153.
24. Leung JM, Levine EH. Left ventricular end-systolic cavity obliteration as an estimate of intraoperative hypovolemia. Anesthesiology 1994;81:1102–1109.
25. Subramaniam B, Talmor D. Echocardiography for management of hypotension in the intensive care unit. Crit Care Med 2007;35:S401–S407.
26. Riesen SC, Doherr MG, Lombard CW. Comparison of Doppler-derived aortic velocities obtained from various transducer sites in healthy dogs and cats. Vet Radiol Ultrasound 2007;48:570–573.
27. Lamia B, Ochagavia A, Monnet X, et al. Echocardiographic prediction of volume responsiveness in critically ill patients with spontaneously breathing activity. Intensive Care Med 2007;33:1125–1132.
28. Cornell CC, Kittleson MD, Della Torre P, et al. Allometric scaling of M-mode cardiac measurements in normal adult dogs. J Vet Intern Med 2004;18:311–321.
29. Boswood A, Haggstrom J, Gordon SG, et al. Effect of Pimobendan in dogs with preclinical myxomatous mitral valve disease and cardiomegaly: the EPIC study—a randomized clinical trial. J Vet Intern Med 2016;30:1765–1779.
30. Teichholz LE, Kreulen T, Herman MV, et al. Problems in echocardiographic volume determinations: echocardiographic-angiographic correlations in the presence of absence of asynergy. Am J Cardiol 1976;37:7–11.
31. Khan KS, Chien PF. Evaluation of a clinical test. I: assessment of reliability. BJOG 2001;108:562–567.
32. Cannesson M, Le Manach Y, Hofer CK, et al. Assessing the diagnostic accuracy of pulse pressure variations for the prediction of fluid responsiveness: a “gray zone” approach. Anesthesiology 2011;115:231–241.
33. Muir WW, Ueyama Y, Pedraza-Toscano A, et al. Arterial blood pressure as a predictor of the response to fluid administration in euvolemic nonhypotensive or hypotensive isoflurane-anesthetized dogs. J Am Vet Med Assoc 2014;245:1021–1027.
34. 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.
35. 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.
36. Fantoni DT, Ida KK, Gimenes AM, et al. Pulse pressure variation as a guide fo r volume expansion in dogs undergoing orthopedic surgery. Vet Anaesth Analg 2017;44:710–718.
37. De Backer D, Pinsky MR. Can one predict fluid responsiveness in spontaneously breathing patients? Intensive Care Med 2007;33:1111–1113.
38. Lewis JF, Kuo LC, Nelson JG, et al. Pulsed Doppler echocardiographic determination of stroke volume and cardiac output: clinical validation of two new methods using the apical window. Circulation 1984;70:425–431.
39. Nguyen HB, Rivers EP, Abrahamian FM, et al. Severe sepsis and septic shock: review of the literature and emergency department management guidelines. Ann Emerg Med 2006;48:28–54.
40. Kittleson MD, Kienle RD. Echocardiography. In: Small animal cardiovascular medicine. St Louis: Mosby, 1998;95–117.
41. Serres F, Chetboul V, Tissier R, et al. Comparison of 3 ultrasound methods for quantifying left ventricular systolic function: correlation with disease severity and prognostic value in dogs with mitral valve disease. J Vet Intern Med 2008;22:566–577.
42. McLean AS. Echocardiography in shock management. Crit Care 2016;20:275.
43. Lichtenstein D. FALLS-protocol: lung ultrasound in hemodynamic assessment of shock. Heart Lung Vessel 2013;5:142–147.
44. Vincent JL, Sakr Y, Sprung CL, et al. Sepsis in European intensive care units: results of the SOAP study. Crit Care Med 2006;34:344–353.
45. Marik PE, Lemson J. Fluid responsiveness: an evolution of our understanding. Br J Anaesth 2014;112:617–620.
46. Miller A, Mandeville J. Predicting and measuring fluid responsiveness with echocardiography. Echo Res Pract 2016;3:G1–G12.
