Copyright is owned by the Author of the thesis. Permission is given for a 
copy to be downloaded by an individual for the purpose of research and 
private study only. The thesis may not be reproduced elsewhere without 
the permission of the Author. 
 



 
 

 

 

PERIOPERATIVE FLUID ADMINISTRATION 

TO OPTIMISE HAEMODYNAMICS 

WITHOUT FLUID OVERLOAD IN ANAESTHETISED DOGS 

 
 
 
 
 

A thesis presented in partial fulfilment 

of the requirements for the degree of  

 
Doctor of Philosophy 

in 

Veterinary Science 

 

at Massey University, Manawatu 

New Zealand 

 

 

HIROKI SANO 

2021 

 
 
 
 
 
 



 ii 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



 iii 

 

 

 

 

 

 

 

Dedication: 

To my colleagues at the veterinary teaching hospital, Massey University, 

Thank you for helping my research projects, 

Thank you for your support and patience, 

I really appreciate your invaluable assistance in this work. 

 

 

 

 

 

 

 

 

 

 

 

 

 



 iv 

ABSTRACT 

Perioperative fluid therapy is the mainstay of anaesthetic management. Fluid 

administration improves haemodynamics during anaesthesia as it increases preload and 

thus cardiac output and blood pressure. However, excessive fluid administration can 

cause detrimental adverse effects, such as haemodulution and oedema, resulting in 

prolonged hospital stay and increased morbidity and mortality in people. Therefore, fluid 

administration should be restricted to those who are able to increase stroke volume or 

cardiac output in response to the fluid administration (responders) and should not be given 

to those who are unable to do so (non-responders) based on the famous “Frank–Starling 

law of the heart” 

 

Previously static parameters such as central venous pressure were believed to be a clinical 

gold standard to estimate preload and fluid responsiveness. Over the last decade, dynamic 

parameters such as pulse pressure variation and pleth variability index have been shown 

to be reliable predictors for fluid responsiveness in people. This study found that pulse 

pressure variation and pleth variability index were more accurate than central venous 

pressure for predicting fluid responsiveness in dogs. 

 

Mini-fluid challenge is another technique that is currently available and can be reliably 

used to determine fluid responsiveness in human medicine. Mini-fluid challenge is an 

administration of a small amount of fluid to increase preload. Thus, fluid responsiveness 

can be assessed based on whether stroke volume increases following mini-fluid challenge 

according to the Frank-Starling curve. The change in stroke volume of a heart at the steep 

portion of the Frank-Starling curve will be greater than at the plateau portion after mini-

fluid challenge. The studies revealed a percentage change in pulse wave transit time (a 



 v 

surrogate parameter of stroke volume, which was also one of results in this thesis) 

following mini-fluid challenge could predict fluid responsiveness in mechanically 

ventilated anaesthetised dogs under an experimental condition, and spontaneously 

breathing anaesthetised dogs undergoing stifle surgery in clinical setting. 

 

Lastly, these methods are still of limited use in veterinary clinical practice because of 

availability of equipment, difficulty of their interpretation and a cumbersome process. 

The main purpose of this thesis was to obtain evidence on how to optimise 

haemodynamics in anaesthetised dogs and prevent excessive fluid administration. The 

time when most practitioners administer a bolus of fluid during anaesthesia is when 

hypotension is encountered because of anaesthesia. Thus, prevention of hypotension 

could avoid excessive fluid administration. Therefore, the study found that prophylactic 

noradrenaline administration, which counteracts some of the cardiovascular adverse 

effects of anaesthesia, was able to prevent hypotension, and thus minimise fluid 

administration in anaesthetised dogs. 

 

Although all of these methods tested in this thesis have pros and cons in clinical veterinary 

practice, they were showen to be able to optimise haemodynamics without fluid overload 

in anaesthetised dogs.  

 

 
 
 
 
 
 
 
 
 
 



 vi 

ACKNOWLEDGEMENT  

I have been fortunate to have many people assist me throughout this research project. 

Firstly, to my supervisors; Associate Professor Paul Chambers, Professor Craig Johnson, 

and Associate professor Nick Cave, the guidance, consultant, routine chat, phone calls, 

and assistance have been greatly appreciated. Working with excellent mentors has been 

an inspiration. I particularly thank you for the mentoring and friendship from Paul, who 

has always been supportive, throughout this journey, despite the challenging environment, 

including intensive clinical duties, teaching responsibility, administrative works and other 

ongoing research projects. I thank you, to Craig, for many chats and discussion about 

many research ideas. I thank you, to Nick for mentoring the writing scientific articles, 

and making study design.  

 

I thank you to all anaesthesia nurses and intern at the teaching hospital, who helped with 

assisting my experiments and collecting clinical cases, even though they had a lot of other 

duties during my research project. 

 

To my colleagues and friends, Drs Mike Gieseg, Paul Wightman, Janis Bridges, Paola 

Giordano, and Vicki Walsh, who have been supportive and offered advice throughout the 

part time PhD and work adventure, thank you for all your advice. To my dear friends Drs. 

Kerrie Lewis and Joana Chagas, who provided all supports throughout this adventure. 

 

To my research collaborators, Drs. Yoshihiro Sugo and Mitsuyoshi Furukawa from Nihon 

Nihon Kohden, who provided a special multiparameter monitor for this project, thank 

you for all supports and unique ideas. 

 



 vii 

Finally, and most importantly, to my family, who without, this work would not have been 

possible. To my parents and grandfather, who without the lifelong support and advice, 

this would not be possible, we only wish my grandfather could be here to read this.  

 

This PhD was partially funded by Massey University Research Funds. All necessary 

approvals were obtained from the Massey University Animal Ethics Committee. 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 



 viii 

LIST OF PUBLICATIONS 
 
Sano H, Seo J, Wightman P, Cave NJ, Gieseg MA, Johnson CB, Chambers P: Evaluation 

of pulse pressure variation and pleth variability index to predict fluid responsiveness in 

mechanically ventilated isoflurane-anesthetized dogs. Journal of Veterinary Emergency 

and Critical Care. 28(4):301-309, 2018. doi: 10.1111/vec.12728. 

 
Sano H & Chambers JP: Ability of pulse wave transit time to detect changes in stroke 

volume and to estimate cardiac output compared to thermodilution technique in 

isoflurane-anaesthetised dogs. Veterinary Anaesthesia and Analgesia. 44(5):1057-1067, 

2017. doi: 10.1016/j.vaa.2016.11.014. 

 
Sano H, Fujiyama M, Wightman P, Cave NJ, Gieseg MA, Johnson CB, Chambers P: 

Investigation of percentage changes in pulse wave transit time induced by mini-fluid 

challenges to predict fluid responsiveness in ventilated dogs. Journal of Veterinary 

Emergency and Critical Care. 29(4):391-398, 2019. doi: 10.1111/vec.12860. 

 
Sano H & Chambers P: Investigation of change in pulse wave transit time following mini-

fluid challenge with colloid predicting fluid responsiveness in spontaneously breathing 

anaesthetised dogs undergoing orthopaedic surgery. 13th World Congress of Veterinary 

Anaesthesiology in Venice, Italy. Poster session, 131, September 2018. The manuscript 

was accepted by Veterinary Anaesthesia and Analgesia in 2020. 

 
Sano H, Chambers P, Bridges J, Johnson C & McGlade K: Effects of noradrenaline 

infusion prior to hypotension on anaesthetic management in dogs undergoing 

ovariohysterectomy. American College of Veterinary Anesthesia and Analgesia Annual 

Conference in Washington D.C., USA. Oral presentation, 111, September 2019. doi: 

10.1016/j.vaa.2019.08.039. The manuscript was accepted by Veterinary Anaesthesia and 

Analgesia in 2020. 

 

 

 

 

 



 ix 

TABLE OF CONTENTS 

 

DEDICATION              iii 

ABSTRACT               iv 

ACKNOWLEDGEMENTS             vi 

LIST OF PUBLICATIONS            vii 

TABLE OF CONTENTS             ix 

LIST OF ABBREVIATIONS           xiii 

LIST OF TABLES              xv 

LIST OF FIGURES           xviii 

INTRODUCTION TO THESIS 

Thesis Structure               1 

Thesis Outline                2 

Chapter 1: LITERATURE REVIEW 

1. Overview of perioperative fluid therapy            5 

2. Cardiac function and fluid loading           13 

3. Assessment of fluid responsiveness          18 

4. Measurement of cardiac output in dogs          29 

5. Perioperative haemodynamic management in dogs        33 

6. Aim and objectives of thesis           38 

7. References              41 

 

Chapter 2: EVALUATION OF PULSE PRESSURE VARIATION AND PLETH 

VARIABILITY INDEX TO PREDICT FLUID RESPONSIVENESS IN 

MECHANICALLY VENTILATED ISOFLURANE-ANAESTHETISED DOGS 

 
Sano H, Seo J, Wightman P, Cave NJ, Gieseg MA, Johnson CB, Chambers P: Evaluation 

of pulse pressure variation and pleth variability index to predict fluid responsiveness in 

mechanically ventilated isoflurane-anesthetized dogs. Journal of Veterinary Emergency 

and Critical Care. 28(4):301-309, 2018. doi: 10.1111/vec.12728. 



 x 

Preface               52 

Abstract              53 

Introduction              55 

Materials and Methods            59 

Results               64 

Discussion              69 

References              77 

 
Chapter 3: ABILITY OF PULSE WAVE TRANSIT TIME TO DETECT CHANGES IN 

STROKE VOLUME AND TO ESTIMATE CARDIAC OUTPUT COMPARED TO 

THERMODILUTION TECHNIQUE IN ISOFLURANE-ANAESTHETISED DOGS 

 
Sano H & Chambers JP: Ability of pulse wave transit time to detect changes in stroke 

volume and to estimate cardiac output compared to thermodilution technique in 

isoflurane-anaesthetised dogs. Veterinary Anaesthesia and Analgesia. 44(5):1057-1067, 

2017. doi: 10.1016/j.vaa.2016.11.014. 

Preface               84 

Abstract              85 

Introduction              87 

Materials and Methods            90 

Results               97 

Discussion            104 

References            109 

 
Chapter 4: INVESTIGATION OF PERCENTAGE CHANGES IN PULSE WAVE 

TRANSIT TIME INDUCED BY MINI-FLUID CHALLENGES TO PREDICT FLUID 

RESPONSIVENESS IN VENTILATED DOGS 

 
Sano H, Fujiyama M, Wightman P, Cave NJ, Gieseg MA, Johnson CB, Chambers P: 

Investigation of percentage changes in pulse wave transit time induced by mini-fluid 

challenges to predict fluid responsiveness in ventilated dogs. Journal of Veterinary 

Emergency and Critical Care. 29(4):391-398, 2019. doi: 10.1111/vec.12860. 

Preface             114 

Abstract            115 

Introduction            117 



 xi 

Materials and Methods          122 

Results             127 

Discussion            132 

References            138 

 
CHAPTER 5: INVESTIGATION OF CHANGE IN PULSE WAVE TRANSIT TIME 

FOLLOWING MINI-FLUID CHALLENGE PREDICTING FLUID 

RESPONSIVENESS IN SPONTANEOUSLY BREATHING ANAESTHETISED 

DOGS 

 
Sano H & Chambers P: Investigation of change in pulse wave transit time following mini-

fluid challenge with colloid predicting fluid responsiveness in spontaneously breathing 

anaesthetised dogs undergoing orthopaedic surgery. 13th World Congress of Veterinary 

Anaesthesiology in Venice, Italy. Poster session, 131, September 2018. The manuscript 

was accepted by Veterinary Anaesthesia and Analgesia in 2020. 

Preface                142 

Abstract            143 

Introduction            146 

Materials and Methods          150 

Results             156 

Discussion            163 

References            169 

 
Chapter 6: EFFECTS OF NORADRENALINE INFUSION PRIOR TO 

HYPOTENSION ON ANAESTHETIC MANAGEMENT IN DOGS UNDERGOING 

OVARIOHYSTERECTOMY 

 
Sano H, Chambers P, Bridges J, Johnson C & McGlade K: Effects of noradrenaline 

infusion prior to hypotension on anaesthetic management in dogs undergoing 

ovariohysterectomy. American College of Veterinary Anesthesia and Analgesia Annual 

Conference in Washington D.C., USA. Oral presentation, 111, September 2019. doi: 

10.1016/j.vaa.2019.08.039. The manuscript was accepted by Veterinary Anaesthesia and 

Analgesia in 2020. 

Preface             174 

Abstract            175 



 xii 

Introduction            177 

Materials and Methods          180 

Results             186 

Discussion            194 

References            200 

 
Chapter 7: GENERAL DISCUSSION AND CONCLUSIONS 

Preface             203 

Principal findings           205 

Discussion            208 

Conclusions            215 

Future Directions           216 

References            220 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



 xiii 

LIST OF ABBREVIATIONS 

ASA  American Society of Anesthesiologists 

ANOVA  analysis of variance  

ASA  American Society of Anesthesiologists 

AUC  area under a curve 

AUROC area under the receiver operating characteristic 

BG  blood glucose (mmol/L) 

CDA  clinical depth of anaesthesia 

CI  cardiac index (mL/kg/minute) 

CO  cardiac output (mL/minute) 

CVP  central venous pressure (mmHg) 

DO2  oxygen delivery (mL/minute) 

ECG  electrocardiogram 

esCO  CO estimated from PWTT (mL) 

esCOIBP CO estimated from PWTT and calibrated with IBP (L/minute) 

esCONIBP CO estimated from PWTT and calibrated with NIBP (L/minute) 

esSV  SV estimated from PWTT (L/minute) 

FE´Iso  end-tidal isoflurane concentration (%) 

FFC  full fluid challenge 

fR  respiratory rate (breaths/minute) 

Hb  haemoglobin (g/dL) 

HR  heart rate (beats/minute) 

IBP  invasive blood pressure (mmHg) 

IM  intramuscularly 

IV  intravenously 



 xiv 

MAP  mean arterial pressure (mmHg) 

MFC  mini-fluid challenge 

NIBP  non-invasive blood pressure (mmHg) 

PCV  packed cell volume (%) 

PE´CO2 end-tidal carbon dioxide tension (mmHg) 

PI  perfusion index 

PIP  peak inspiratory pressure (cmH2O) 

PPV  pulse pressure variation (%) 

PVI  pleth variability index (%) 

PWTT  pulse wave transit time (msecond) 

ΔPWTT percentage change in PWTT (%) 

ΔPWTT1,2,3,10 ΔPWTT after 1, 2, 3, and 10 mL/kg colloid administration (%) 

ΔPWTTFFC percentage change in PWTT over FFC (%) 

ΔPWTTMFC percentage change in PWTT over MFC (%) 

ROC  receiver operator characteristic 

SD  standard deviation 

SE  standard error 

SV  stroke volume (mL) 

SVV  stroke volume variation (%) 

SVR  systemic vascular resistance (dynes/second/cm5) 

ΔSV  percentage change in SV (%) 

T  temperature (℃) 

TEE  transoesophageal echocardiography 

TDCO  CO measured by thermodilution technique (L/minute)  

TDSV  SV measured by thermodilution technique (mL) 



 xv 

TS  total solids (g/L) 

VT  tidal volume (mL/kg) 

VTI  velocity time integral (cm) 

ΔVTI  percentage change in VTI (%) 

ΔVTI1,2,3,10 ΔVTI after 1, 2, 3, and 10 mL/kg colloid administration (%) 

ΔVTIFFC  percentage change in VTI over FFC (%) 

ΔVTIMFC percentage change in VTI over MFC (%) 

95% CI 95% confidence interval 

ΔPWTT percentage change in PWTT (%) 

ΔPWTTFFC percentage change in PWTT over FFC (%) 

ΔPWTTMFC percentage change in PWTT over MFC (%) 

ΔPWTT1,2,3,10 ΔPWTT after 1, 2, 3, and 10 mL/kg colloid administration (%) 

ΔSV  percentage change in SV (%) 

ΔVTI  percentage change in VTI (%) 

ΔVTIFFC percentage change in VTI over FFC (%) 

ΔVTIMFC percentage change in VTI over MFC (%) 

ΔVTI1,2,3,10 ΔVTI after 1, 2, 3, and 10 mL/kg colloid administration (%) 

 

 

 

 

 

 

 

 

 

 

 



 xvi 

LIST OF TABLES 

Chapter 2 

 
Table 2.1 Comparisons of cardiorespiratory variables before and after the fluid challenge 

between responders and non-responders. Data are mean (standard deviation). *Significant 

difference between responders and non-responders (p < 0.05). CVP, central venous 

pressure; VTI, velocity time integral; PPV, pulse pressure variation; PVI, pleth variability 

index; PIP, peak inspiratory pressure; PE´CO2, End-tidal partial pressure of carbon 

dioxide; FE´Iso, end-tidal concentration of isoflurane. 

 
Table 2.2 Comparisons of percentage changes in cardiorespiratory variables after the 

fluid challenge between responders and non-responders. Data are mean (standard 

deviation). *Significant difference between responders and non-responders (p < 0.05). 

CVP, central venous pressure; VTI, velocity time integral; PPV, pulse pressure variation; 

PVI, pleth variability index; PIP, peak inspiratory pressure; PE´CO2, end-tidal partial 

pressure of carbon dioxide; FE´Iso, end-tidal concentration of isoflurane. 

 
Table 2.3 Areas under the receiver operator characteristic (ROC) curves and cutoff values 

for the prediction of fluid responsiveness. AUC: Area under the curve, 95%CI: 95% 

confidence interval, CVP: Central venous pressure, PPV: Pulse pressure variation, PVI: 

Pleth variability index, Sen: Sensitivity, Spe: Specificity. 

 

Chapter 3 

 

Table 3.1 Cardiovascular parameters recorded in eight isoflurane-anaesthetised dogs 

before drug administration (baseline1: BL1), during phenylephrine administration (0.5 

μg/kg/minute: PH1; 1 μg/kg/minute: PH2), 30 minutes after cessation of PH1 (baseline2: 

BL2), during high isoflurane administration (2.0-2.5%: ISO1; 2.5-3.0%: ISO2), 30 

minutes after cessation of ISO2 (baseline3: BL3), during dobutamine administration (1 

μg/kg/minute: DO1; 2 μg/kg/minute: DO2) and 30 minutes after cessation of DO2 

(baseline4: BL4) 

 

Chapter 4 

 

Table 4.1 Cardiovascular variables before and after the mini-fluid challenges (1, 2 and 3 



 xvii 

mL/kg) and the fluid challenge (10 mL/kg) in responders and non-responders. Data are 

mean (standard deviation). HR; heat rate, MAP; mean arterial pressure, CVP; central 

venous pressure, VTI; velocity time integral, PWTT; pulse wave transit time. Some 

variables are extracted from the concurrent study (Chapter 2)(Sano et al., 2018). *p < 

0.05 between responders (n = 14) and non-responders (n = 10). †p < 0.05 between baseline 

and either 1, 2, 3 or 10 mL/kg.  

 
Table 4.2 Comparison of ΔVTI1,2,3,10 and ΔPWTT1,2,3,10 after the mini-fluid challenges (1, 

2 and 3 mL/kg) and the fluid challenge (10 mL/kg) in responders and non-responders. 

Data are mean (standard deviation). ΔVTI1,2,3,10, percentage change in velocity time 

integral of main pulmonary arterial blood flow after 1, 2, 3, and 10 mL/kg of fluid; 

ΔPWTT1,2,3,10, percentage change in pulse wave transit time after 1, 2, 3, and 10 mL/kg 

of fluid. Some variables are extracted from the concurrent study (Chapter 2)(Sano et al., 

2018). *p < 0.05 between responders (n = 14) and non-responders (n = 10). 

Table 4.3 Areas under the receiver operator characteristic (ROC) curves and cutoff values 

for the prediction of fluid responsiveness. AUC: Area under the curve, 95%CI: 95% 

confidence interval, ΔVTI1,2,3,10, percentage change in velocity time integral of main 

pulmonary arterial blood flow after 1, 2, 3, and 10 mL/kg of fluid; ΔPWTT1,2,3,10, 

percentage change in pulse wave transit time after 1, 2, 3, and 10 mL/kg of fluid. *p < 

0.05 between AUC and 0.5. 

Chapter 5 

Table 5.1 Characteristics of 45 dogs [responders (n = 19) and non-responders (n = 26)] 

before anaesthesia. Data are presented as mean ± standard deviation if the data were 

normally distributed or median (range) if not. *p<0.05 between responders (n = 19) and 

non-responders (n = 26). fR, respiratory rate; BUN, blood urea nitrogen; TTA, Tibial 

tuberosity advancement; TPLO, Tibial plateau leveling osteotomy; NA, non-applicable 

because BUN was 5-15 mg/dL for all dogs. 

 

Table 5.2 Cardiorespiratory variables before and after the mini-fluid challenges (MFC: 3 

mL/kg) and the full fluid challenges (FFC: 6 mL/kg) in responders and non-responders. 

Data are presented as median (interquartile range). *p < 0.017 compared to baseline using 

Wilcoxon test with Bonferroni adjustment. †p < 0.05 between responders (n = 19) and 



 xviii 

non-responders (n = 26). HR, heart rate; MAP, mean arterial pressure; VTI, velocity time 

integral; PWTT, pulse wave transit time; fR, respiratory rate; PE´CO2, partial pressure of 

end-tidal carbon dioxide; FE´Iso, end-tidal isoflurane concentration. 

 

Table 5.3 Comparison of percentage changes in velocity time integral (VTI) and pulse 

wave transit time (PWTT) over the mini-fluid challenges (MFC: 3 mL/kg) and the full 

fluid challenges (FFC: 6 mL/kg) in responders and non-responders. Data are presented as 

mean ± standard deviation. *p < 0.05 between responders (n = 19) and non-responders (n 

= 26). ΔHRMFC, ΔMAPMFC, ΔVTIMFC and ΔPWTTMFC, percentage change in heart rate, 

mean arterial pressure, VTI and PWTT over MFC; ΔHRFFC, ΔMAPFFC, ΔVTIFFC and 

ΔPWTTFFC, percentage change in heart rate, mean arterial pressure, VTI and PWTT over 

FFC; NA, non-applicable because the definition of the responders was 15% > ΔVTIFFC. 

 

Table 5.4 Ability of percentage change in velocity time integral (ΔVTIMFC) and pulse 

wave transit time (ΔPWTTMFC) over the mini-fluid challenge to predict increase in stroke 

volume of 15%. AUROC, area under the receiver operating characteristic curve; 95%CI, 

95% confidence interval, Sen; sensitivity, Spe; specificity. 

 

Chapter 6 

Table 6.1 Preoperative characteristics of dogs between noradrenaline and control groups. 

Breed is expressed as breed name (number of dogs). Other values are medians (range). 

BCS, body condition score; HR, heart rate; fR, respiratory rate; T, temperature; PCV, 

packed cell volume; TS, total solids; BG, blood glucose; BUN, blood urea nitrogen. BUN 

was 5-15 mg/dL for all dogs. 

 
Table 6.2 Intraoperative values from dogs undergoing ovariohysterectomy between 

noradrenaline and control groups. Values are medians (range). HR, heart rate; MAP, mean 

arterial pressure; CI, cardiac index; fR, respiratory rate; PE´CO2, partial pressure of end-

tidal carbon dioxide; FE´Iso, End-tidal isoflurane concentration; CDA, Clinical depth of 

anaesthesia; T, temperature. 

 

 

 



 xix 

LIST OF FIGURES 

Chapter 1 

 
Figure 1.1 Avoidance of both hypo- and hypervolaemia is the aim of intraoperative fluid 

therapy in order to prevent adverse outcomes. Modified from Doherty & Buggy (Doherty 

& Buggy, 2012) 

 
Figure 1.2 Frank-Starling curve. An increase in preload increase in stroke volume and 

cardiac output. Figure b showed the curve when the contractility is increased or the 

afterload is reduced, while the curve of Figure c is depicted when the contractility is 

decreased, or the afterload is increased. However, the excessive preload overstretches the 

myocyte, resulting in decreased or unchanged stroke volume. 

 

Figure 1.3 Fluid responder and non-responder. On the steep portion of the curve, changes 

in preload will result in large changes in stroke volume. A patient on the steep portion of 

the curve would have an increased stroke volume in response to fluid loading. Thus, this 

patient is a fluid responder. On the plateau portion of the curve, identical changes in 

preload will not alter stroke volume as much as they would for a patient on the steep 

portion of the curve. Hence, this patient is a fluid non-responder. 

 

Figure 1.4 Stroke volume variation (SVV) A patient on the steep portion of the curve 

would have an increased SV in response to fluid loading. Thus, the patient with large 

SVV is a fluid responder. On the plateau portion of the curve, same changes in preload 

will not alter SV as much as they would for a patient on the steep portion of the curve. 

Hence, the patient with small SVV is a fluid non-responder. 

 
Figure 1.5 Patients A and B have different curves because of different contractility and 

afterload. Given a same change in absolute preload, patient A has a greater change in SV 

than does patient B. Patient A is a responder and therefore has a greater variation in SV 

even though absolute changes in preload are identical for these two patients.  

 

Figure 1.6 Pulse pressure variation (PPV) PPV, 100 × (PPMax - PPMin)/Average PP; 

Average PP, (PPMax + PPMin)/2; PA, arterial pressure; PAW, airway pressure; PPMax, 



 xx 

maximum pulse pressure after a positive pressure breath; PPMin, minimum pulse 

pressure after a positive pressure breath (from Scott et al. 2001).  

 
Figure 1.7 Pleth Variability Index (PVI) PVI, 100 × (PImax - PImin)/PImax; PImax, Maximum 

perfusion index; PImin, Minimum PI (from Masimo HP; https://www.masimo.com/pvi/) 

 
Figure 1.8 The Frank-Starling curve of the heart. Mini-fluid challenge is a strategy to 

assess fluid responsiveness based on a change in stroke volume (SV) after a small loading 

dose of fluid. The change in SV at the steep portion of the curve (responders) will be 

greater than at the plateau portion (non-responders) after both the mini-fluid challenge 

and fluid challenge. Therefore, the magnitude of the change in SV after the mini-fluid 

challenge could predict responsiveness to the fluid challenge. 

 
Figure 1.9 Pulse Wave Transit Time (PWTT). PWTT was calculated as the time from the 

ECG R-wave peak to the rise point of the pulse oximeter wave. The rise point of the pulse 

wave was defined as the point at which the pulse wave reached 30% of its peak amplitude. 

 
Figure 1.10 Cerebral blood flow autoregulation (Paulson et al., 1990). It typically 

operates between MAP of the order of 60 and 150 mmHg (normal MAP). 

 
Figure 1.11 Renal perfusion flow autoregulation (Shipley & Study, 1951). Renal 

perfusion flow (RPF) drops off at 70 mmHg but the plateau really starts at about 110 

mmHg. Glomerular filtration (GFR) plateaus at 120 mmHg and by the time renal artery 

mean blood pressure is 70 mmHg the GFR is at about 70%. 

 

Chapter 2 

 

Figure 2.1 The Frank–Starling curve of the heart (A), definition of Pulse pressure 

variation (PPV) (B) and Pleth variability index (PVI) (C). In mechanically ventilated 

patients, stroke volume (SV) can change greater (a) in response to the variation of preload 

(b) caused by positive-pressure ventilation (Responders). In contrast, an increase in SV 

is relatively insensitive (c) to preload changes (d) on the plateau portion of the curve 



 xxi 

(Non-responders). PP, pulse pressure (systolic arterial pressure - diastolic arterial 

pressure); Average PP, (PPmax + PPmin)/2; PI, perfusion index; max, maximum; min, 

minimum.  

 
Figure 2.2 Comparison of areas under the receiver operator characteristic (ROC) curve 

for CVP, PPV and PVI. CVP, central venous pressure; PPV, pulse pressure variation; PVI, 

pleth variability index. 

 

Chapter 3 

Figure 3.1 Pulse wave transit time (PWTT). PWTT was calculated as the time from the 

electrocardiogram R-wave peak to the rise point of the pulse oximeter wave. The rise 

point of the pulse wave was defined as the point at which the pulse wave reached 30% of 

its peak amplitude. 

Figure 3.2 Timeline of the study. After the calibration, 10 measurements were taken; 

before drug administration (baseline1: BL1), during phenylephrine administration (0.5 

μg/kg/minute: PH1; 1 μg/kg/minute: PH2), 30 minutes after cessation of PH1 (baseline2: 

BL2), during high isoflurane administration (2.0-2.5%: ISO1; 2.5-3.0%: ISO2), 30 

minutes after cessation of ISO2 (baseline3: BL3), during dobutamine administration (1 

μg/kg/minute: DO1; 2 μg/kg/minute: DO2) and 30 minutes after cessation of DO2 

(baseline4: BL4).  

Figure 3.3 Area under the receiver operator characteristic (ROC) curve for the percentage 

change in pulse wave transit time (PWTT) to detect a 15% change in stroke volume 

derived from the thermodilution technique (TDSV). The closer to 1.0 of an area under the 

ROC curve indicates, the more reliable diagnostic method. 95% CI, 95% confidence 

interval. The area under the ROC curve and the cut-off value of the percentage change in 

PWTT to detect 15% change in stroke volume were 0.91 (95% CI, 0.85e0.98; p < 0.001) 

and 2.7% (sensitivity: 86%; specificity: 81%), respectively.  

Figure 3.4 Four-quadrant plot with linear regression analysis between percentage 

changes in PWTT and TDSV. The dotted lines limit an exclusion zone of ±15% for TDSV 

and 2.7% for PWTT. PWTT, pulse wave transit time; TDSV, stroke volume measured by 



 xxii 

the thermodilution technique.  

Figure 3.5 Bland-Altman plots for the difference between the methods plotted against 

their mean. (A) TDCO versus esCO with NIBP calibration, (B) TDCO versus esCO with 

IBP calibration. The solid line represents the mean bias, the two broken lines indicate the 

mean bias ± 1.96 standard deviation; the dotted line is the line of equality. esCO, 

estimated cardiac output based on pulse wave transit time; IBP, invasive blood pressure; 

NIBP, non-invasive blood pressure; SD, standard deviation; TDCO, cardiac output 

measured by the thermodilution technique.  

 

Chapter 4 

Figure 4.1 The Frank-Starling curve of the heart. Mini-fluid challenge is a strategy to 

assess fluid responsiveness based on a change in stroke volume (SV) after a small loading 

dose of fluid. The change in SV at the steep portion of the curve (responders) will be 

greater than at the plateau portion (non-responders) after both the mini-fluid challenge 

and fluid challenge. Therefore, the magnitude of the change in SV after the mini-fluid 

challenge could predict responsiveness to the fluid challenge. 

 

Figure 4.2 Correlation between ΔPWTT3 and ΔVTI10,  ΔVTI3, and ΔPWTT10 and 

ΔVTI10 ΔVTI3,10, percentage change in velocity time integral of main pulmonary arterial 

blood flow after 3, and 10 mL/kg of fluid; ΔPWTT3,10, percentage change in pulse wave 

transit time after 3, and 10 mL/kg of fluid. *p < 0.05 

 
Figure 4.3 Receiver operator characteristic (ROC) curve for ΔVTI3 and ΔPWTT3. ΔVTI3 

and ΔPWTT3, percentage changes in velocity time integral of main pulmonary arterial 

blood flow and pulse wave transit time after a 3 mL/kg mini-fluid challenge; AUC, area 

under the curve; 95%CI, 95% confidence interval. 

 
Figure 4.3 The grey zone and the best cutoff value within the optimal cutoff values for 

ΔPWTT3. Distribution of the cutoffs for each bootstrapped population (1000 “optimal” 

values). Grey rectangle (the grey zone), 95% confidence interval for the optimal cutoffs; 

Black line (the best cutoff value), mean; ΔPWTT3, percentage changes in pulse wave 

transit time after a 3 mL/kg mini-fluid challenge. 



 xxiii 

 

Chapter 5 

Figure 5.1 The Frank-Starling curve of the heart. Mini-fluid challenge is a strategy to 

assess fluid responsiveness based on a change in stroke volume (SV) after a small loading 

dose of fluid. The change in SV at the steep portion of the curve (responders) will be 

greater than at the plateau portion (non-responders) after both the mini-fluid challenge 

and fluid challenge. Therefore, the magnitude of the change in SV after the mini-fluid 

challenge could predict responsiveness to the fluid challenge. 

 
Figure 5.2 Pulse Wave Transit Time (PWTT). PWTT was calculated as the time from the 

ECG R-wave peak to the rise point of the pulse oximeter wave. The rise point of the pulse 

wave was defined as the point at which the pulse wave reached 30% of its peak amplitude. 

 
Figure 5.3 Receiver operator characteristic curve for ΔVTIMFC and ΔPWTTMFC. 

ΔVTIMFC and ΔPWTTMFC, percentage changes in velocity time integral of aortic blood 

flow and pulse wave transit time after a 3 mL/kg mini-fluid challenge. 

 
Figure 5.4 The grey zone and the best cutoff value within the optimal values for ΔVTIMFC 

and ΔPWTTMFC. Distribution of the cutoffs for each bootstrapped population (1000 

“optimal” values). Grey rectangle, grey zone; Black line (the best cutoff value); ΔVTIMFC 

and ΔPWTTMFC, percentage changes in velocity time integral of aortic blood flow and 

pulse wave transit time after a 3 mL/kg of mini-fluid challenge. 

 

Chapter 6 

Figure 6.1 Mean ± SE haemodynamic parameters over the time points. T0, 5 minutes 

after the induction; T1, 10 minutes after the infusion started; T2 ,before start of the 

surgery; T3 ,time of skin incision; T4 ,time of removal of the first ovary; T5 ,time of 

removal of the second ovary; T6 ,time of closure of the abdominal wall; T7 ,immediately 

after the skin was sutured. Overall, mean arterial pressure was significantly greater (p < 

0.01) and heart rate was significantly lower (p < 0.01) in the noradrenaline group than 

those in the control group, while cardiac index was no significant between groups (p = 

0.12). 

 
Figure 6.2 Mean ± SE end-tidal isoflurane concentration (Left) and frequency of clinical 



 xxiv 

depth of anaesthesia (CDA) scores (Right) over the time points. T0, 5 minutes after the 

induction; T1, 10 minutes after the infusion started; T2 ,before start of the surgery; 

T3 ,time of skin incision; T4 ,time of removal of the first ovary; T5 ,time of removal of 

the second ovary; T6 ,time of closure of the abdominal wall; T7 ,immediately after the 

skin was sutured. CDA scores of 1, 2 and 3 indicated light anaesthesia, adequate surgical 

anaesthesia and an excessive depth of anaesthesia, respectively. C, control group; N. 

norepinephrine group. End-tidal isoflurane concentration in the noradrenaline group was 

significantly higher than that in the control group (p < 0.01) although to account for 

repeated measures in the CDA data, there was no effect of treatment with a cumulative 

link mixed model (p = 0.66). 

 
Figure 6.3 Mean ± SE haematology parameters before and after the surgery. Blood 

samples were taken from the arterial catheter immediately after the catheter was placed 

(Pre) and immediately after the end of anaesthesia (Post). Packed cell volume (p < 0.001), 

total solids (p < 0.001) and blood glucose (p = 0.023) were significantly lower in the 

control group than those in the noradrenaline group after the surgery, while there was no 

effect of treatment on lactate (p = 0.116) and creatinine between groups (p = 0.06). 

 

Chapter 7 

Figure 7.1 Clinical interpretation of haemodynamics. MAP can be calculated by CO and 

SVR, and CO can be calculated by HR and SV.  SV is determined by preload, contractility 

and afterload. MAP and HR is only parameters that can be obtained clinically, which are 

not enough to comprehend haemodynamics. MAP; mean arterial pressure, CO; cardiac 

output, SVR; systemic vascular resistance, HR; heart rate, SV; stroke volume. 

 

 

 

 

 

 

 

 

 
 



Introduction 
 

 1 

 
INTRODUCTION 

 
 

THESIS STRUCTURE 

The studies presented in this thesis are in form of manuscripts published in peer-reviewed 

journals and presented at major anaesthesia conferences, and formatted for the style of 

the journal they were published in. Consequently, there is some repetition of background 

information and methods in some of the chapters, and also the units (SI or American) 

depends on the style of the journal. All manuscripts have been standardised to one 

referencing style throughout the thesis. References are included at the end of each chapter. 

Throughout Chapter 1-6, figures and tables are labelled as Figure 1. Table 1 etc, in line 

with the published manuscript. As such, there are multiple Figure 1s etc throughout the 

thesis, however, where required, each figure is clearly identified as Chapter 2, Figure 

2.1 etc.  

 
 
 
 
 

 
 
 
 

 
 
 
 

 



Introduction  
 

 2 

THESIS OUTLINE 

Chapter 1 of the thesis is a literature review, which provides the reader with a brief 

overview of perioperative fluid therapy and cardiac function with fluid loading, then 

moves on to assessment of fluid responsiveness. An overview of measurement of cardiac 

output is covered, as well as perioperative haemodynamic management in dogs. The 

literature review reveals that there are many questions still unanswered in regard to 

perioperative fluid administration to optimise haemodynamics without fluid overload in 

anaesthetised dogs. 

 

The aim of the study described in Chapter 2 is to investigate whether dynamic indices 

(pulse pressure variation and pleth variability index) could predict fluid responsiveness 

more accurately than classical static index (central venous pressure) in dogs. These 

dynamic indices have been proved to be reliable predictors for fluid responsiveness in 

people but have not been investigated in dogs at the time of the study. However, the use 

of these in clinical veterinary practice is impractical for various reasons. 

 

Chapter 3 evaluated the feasibility of pulse wave transit time as a surrogate of stroke 

volume in dogs. Particularly, the study focused on the ability of change in pulse wave 

transit time to detect changes in actual stroke volume. Accurate measurement of change 

in stroke volume is necessary for an alternative technique to assess fluid responsiveness, 

the mini-fluid challenge studied in Chapter 4 and 5. 

 

Chapter 4 and 5 investigated whether a percentage change in pulse wave transit time (a 

surrogate parameter of stroke volume) following mini-fluid challenge could predict fluid 

responsiveness in mechanically ventilated anaesthetised dogs under experimental 

conditions (Chapter 4), and spontaneously breathing anaesthetised dogs undergoing 



Introduction 
 

 3 

stifle surgery in clinical setting (Chapter 5). 

 

Chapter 6 investigated whether prophylactic noradrenaline administration would be able 

to prevent hypotension, and thus minimise the requirement for fluid administration in 

anaesthetised dogs. This study was performed because all methods tested in previous 

chapters are still of limited use in veterinary clinical practice due to availability of 

equipment, difficulty of their interpretation, and a cumbersome process. 

 

To conclude, in the final chapter of this thesis, a discussion of the key findings of this 

research is provided, including future direction. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



Introduction  
 

 4 

 

 

 



Chapter 1 

 5 

 

Chapter 1 

 

Literature review 

 

1. OVERVIEW OF PERIOPERATIVE FLUID THERAPY 

1.1 Introduction 

Fluid therapy has had an important role of haemodynamic management since successful 

treatment of cholera patients with intravenous fluid was first reported by O’Shaughnessy 

and Latta in the 1830s (Latta, 1832; O’Shaughnessy, 1831). Fluid administration 

compensated for the luminal losses, and corrected the circulation deficit. In the 1880s, 

fluid therapy was established as a treatment of haemorrhage and shock, after it was shown 

to improve the circulation of a patient suffering from antepartum haemorrhage (Egerton 

Jennings, 1882; Thomas, 1898). Perioperative continuous fluid administration was 

described in 1924 as restoring the ongoing fluid loss, and supporting the circulation in 

surgical patients (Matas, 1924), and it was found to prevent pre-renal injury, resulting in 

improved morbidity and mortality (Coller, Dick, & Maddock, 1936). Therefore, the 

efficacy of perioperative fluid therapy to support patient care has been established for 

many years. 

 

In the late 1900s, the concept of supranormal fluid resuscitation was proposed by 

Shoemaker et al (Shoemaker, Appel, Kram, Waxman, & Lee, 1988). They observed that 



Chapter 1 

 6 

survivors from shock who received fluid resuscitation had supranormal levels of oxygen 

delivery and cardiac output compared to non-survivors who did not. The concept that 

early supranormal fluid resuscitation improved the outcome was supported by other 

(Bishop et al., 1995; Bishop et al., 1993). However, several prospective randomised 

clinical studies showed conflicting results (Boyd, Grounds, & Bennett, 1993; Durham, 

Neunaber, Mazuski, Shapiro, & Baue, 1996; Gattinoni et al., 1995). In addition, although 

Shoemaker et al. showed a benefit of fluid therapy in resuscitating patients with shock, 

they also concluded that fluid therapy was also a contributing factor to the outcome of 

critically injured patients. Patients who could achieve supranormal haemodynamic values 

were more likely to survive than those who could not, regardless of the resuscitation 

technique (Velmahos et al., 2000). Therefore, the need for aggressive fluid therapy to 

achieve supranormal resuscitation has been reconsidered. 

 

Nowadays, perioperative fluid therapy is commonly used to provide essential fluid for 

maintenance of body functions, and replace fluid losses due to co-existing diseases, 

haemorrhage, evaporation from surgical sites, and accumulation in an unknown third 

space. Perioperative fluid is routinely administered to achieve supranormal 

haemodynamic optimisation. However, more recently, the focus has been focused on 

perioperative fluid restriction compared to liberal fluid therapy.  Lobo et al. demonstrated 

that restricting perioperative intravenous fluid to less than traditionally administered, 

increased gastric emptying and decreased complications and length of hospital stay in 20 

patients undergoing elective colonic resection (Lobo et al., 2002), compared to patients 

given liberal fluid therapy. Prospective randomised trials and a meta-analysis showed the 

benefits of fluid restriction and avoidance of fluid overload in improving clinical 

outcomes in different clinical settings (Brandstrup et al., 2003; McArdle et al., 2009; 



Chapter 1 

 7 

Nisanevich et al., 2005; Rahbari et al., 2009). Although the aim of fluid administration is 

to support adequate tissue perfusion, excessive fluid administration caused tissue oedema 

(Holte, Jensen, & Kehlet, 2003), and impaired oxygen and nutrient delivery to tissues 

(Cotton, Guy, Morris, & Abumrad, 2006).  

 

Therefore, fluid imbalance, a term to describe either too much or too little fluid, should 

be avoided, which is challenging in clinical practice. 

 

1.2 Traditional perioperative fluid therapy 

The purpose of perioperative fluid therapy is to maintain sufficient circulating volume to 

ensure perfusion of the organs and oxygen delivery to the tissues. Patients might become 

hypovolaemic during anaesthesia and surgery because of fasting overnight, inability to 

drink water due to the stressful environment such as a hospital ward, ongoing losses from 

unexpected urinary output due to underlying diseases, blood loss, and evaporation from 

surgical sites or expired gas. Thus, large volumes of crystalloid were traditionally 

administered to replace the fluid deficit during anaesthesia. 

 

A fluid bolus is also commonly administered when hypotension induced by general 

anaesthetic drugs occurs. General anaesthesia causes vasodilatation, leading to relative 

hypovolaemia (Noel-Morgan & Muir, 2018).Thus, since cardiac output is dependent on 

filling pressure, it is considered rational to give fluids to correct the relative hypovolaemia 

and increase filling pressure. However, it has been shown that fluid loading has little or 

no influence on anaesthesia-related hypotension in people (Norberg et al., 2007) and dogs 

(Valverde, Gianotti, Rioja-Garcia, & Hathway, 2012). This is probably due to the negative 

inotropic effect of anaesthesia, which prevents an increase in cardiac output in response 



Chapter 1 

 8 

to an increased filling pressure. Thus, hypotension induced by anaesthesia may be more 

appropriately treated using inotropic and vasopressor drugs, instead of a large volume of 

fluid, which may case patients to become hypervolaemic. 

 

Furthermore, it was found that during major surgical procedures, patients experience a 

marked loss of extravascular fluid regardless of blood loss. Following from that, an 

implausible hypothesis was posted in the 1960s (Shires, Williams, & Brown, 1961), 

which was that body cavity exposure during surgery induced an internal redistribution of 

extracellular fluid to a “third space” resulting in a reduction in the circulating volume. 

This third space into which fluid disappeared was hypothesised to be in traumatised tissue 

or the gastrointestinal tract.  To compensate for the loss of circulating volume due to this 

hypothetical third space (Jacob, Chappell, & Rehm, 2009), large volumes of crystalloid 

fluids were historically administered perioperatively. However, with the use of methods 

to measure extracellular volume with radioactive tracers, recent trials found an unchanged 

or even increased extracellular volume during surgery, the opposite of the previous third 

space concept. Thus, traditional large volumes of fluid therapy cause fluid overload, 

which may result in haemodilution (including dilution of clotting factors), interstitial 

oedema and thus impaired oxygen and nutrition delivery. Indeed, excessive fluid 

resuscitation has been reported to cause oedema, with increased morbidity (Arieff, 1999), 

impaired coagulation (Ruttmann, James, & Aronson, 1998), bacterial translocation and 

sepsis (Ratner, Lysenko, Paul, & Weiser, 2005) and poor wound healing (Lang, Boldt, 

Suttner, & Haisch, 2001). In addition, the increase in body weight from excess fluid has 

been shown to be correlated with postoperative morbidity and mortality (Holte et al., 

2003; Lowell, Schifferdecker, Driscoll, Benotti, & Bistrian, 1990). Subsequently, a 

systematic review published in 2006 concluded that the concept of third space was 



Chapter 1 

 9 

fundamentally incorrect (Brandstrup, Svensen, & Engquist, 2006). Therefore, restrictive 

perioperative fluid therapy started to be investigated. 

 

1.3 Restrictive perioperative fluid therapy 

Early randomised trials supported positive benefits with restrictive fluid therapy in 

abdominal surgery, with a faster return of gastrointestinal function, fewer complications 

and shorter hospital stay compared to the traditional, or  liberal fluid therapy (Brandstrup 

et al., 2003; Lobo et al., 2002; Nisanevich et al., 2005). For example, in a randomised, 

multicentre trial comparing traditional and restrictive fluid groups in 141 patients 

undergoing colorectal surgery, patients in the restrictive group had a lower complication 

rate (anastomotic leakage, wound infection, and cardiovascular and pulmonary 

compromises) in the restrictive group compared to the liberal group (33% vs 51%, p = 

0.02) (Brandstrup et al., 2003). However, there were also studies that have shown no 

benefit from a restrictive fluid protocol compared with the traditional protocol (MacKay 

et al., 2006; Vermeulen, Hofland, Legemate, & Ubbink, 2009). A review of postoperative 

outcome from restrictive vs traditional fluid therapy in seven randomised trials (six 

involving major abdominal surgery and one involving knee arthroplasty), revealed that 

three studies found an improved outcome (faster return of gastrointestinal function and 

reduced hospital length of stay) with a restrictive fluid regimen, whereas two studies 

found no difference, and two studies found that although the total number of 

complications was reduced, the number of patients with complications was not 

significantly reduced (Bundgaard-Nielsen, Secher, & Kehlet, 2009). However, it is 

difficult to form a holistic interpretation of the studies, because there was variation in the 

definition of liberal or restrictive protocols in clinical practice, whereby a restrictive 

regime in one centre was actually liberal in another. The studies also varied in design, 



Chapter 1 

 10 

types of fluid administered, indications for administering additional fluid, outcomes 

variables, and definitions of intra- and postoperative periods (Corcoran, Rhodes, Clarke, 

Myles, & Ho, 2012; P. S. Myles et al., 2018; Rahbari et al., 2009). Eventually, the 

conclusion was that optimal fluid therapy requires the assessment of each individual’s 

haemodynamic status, to administer what is called “individualised goal-directed fluid 

therapy” (Chong, Wang, Berbenetz, & McConachie, 2018; Xu et al., 2018). 

 

Outcomes may be improved if fluid therapy is individualised based on objective feedback 

on the patient’s individual fluid responsiveness. The “fluid responsiveness” is derived 

from the physiological principal of the Frank-Starling curve and determined by dynamic 

predictors, described later in this Chapter (see Section 3. ASSESSMENT OF FLUID 

RESPONSIVENESS in Chapter 1”). A meta-analysis of ninety-five randomised trails 

concluded that individualised goal-directed fluid therapy using dynamic predictors 

reduced postoperative morbidity and mortality in adult surgical patients compared to 

conventional fluid therapy although the quality of evidence was low (Chong et al., 2018). 

Another meta-analysis of eleven randomised studies showed that gastrointestinal function 

improved with goal-directed fluid therapy compared with conventional fluid therapy in 

patients undergoing colorectal surgery (Xu et al., 2018). Both meta-analyses concluded 

that individualised goal-directed fluid therapy reduced perioperative fluid administration 

compared to historical fluid therapy. Thus, more recent expert guideline/consensus 

statements/meta-analysis on perioperative fluid therapy have supported more restrictive 

fluid regimens (Jia et al., 2017; Lassen et al., 2009; P. Myles et al., 2017; Schol, Terink, 

Lance, & Scheepers, 2016; Self et al., 2018). Individualised goal-directed fluid therapy 

based on fluid responsiveness derived from the Frank-Starling curve has yet to be widely 

accepted in veterinary practice, but it is argued that it should be, because there is no reason 



Chapter 1 

 11 

to believe there are fundamental species differences in this respect, and both 

hypovolaemia and hypervolaemia are known to cause increased perioperative morbidity 

and mortality (Doherty & Buggy, 2012; Navarro et al., 2015) (Figure 1.1). 

 

 

Figure 1 Avoidance of both hypo- and hypervolaemia is the aim of intraoperative fluid 

therapy in order to prevent adverse outcomes. Modified from Doherty & Buggy (Doherty 

& Buggy, 2012) 

 

 

1.4 Fluid Therapy Guidelines for Dogs and Cats 

The basic concept of fluid therapy for dogs may be extrapolated from the human literature 

because physiology of mammals is similar. American Animal Hospital Association 

(AAHA)/American Association of Feline Practitioners (AAFP) published fluid therapy 

guidelines (Davis et al., 2013), anaesthesia guidelines (Bednarski et al., 2011) and 

monitoring guidelines (Grubb et al., 2020) including perioperative fluid therapy for dogs. 

These guidelines are recommendations from AAHA/ AAFP based on expert opinion 

rather than controlled trials. The guidelines recommend therapy that is tailored to each 



Chapter 1 

 12 

patient and constantly re-evaluated and reformulated according to changes in status. 

During the perioperative period, the proposed benefits of fluid therapy include 

replacement of normal ongoing fluid losses and fluid losses related to surgery, and 

support of cardiovascular function based on the Frank-Starling curve. Current 

recommendations are to deliver < 10 mL/kg/hr to avoid adverse effects associated with 

hypervolaemia, particularly in cats due to their smaller blood volume to weight ratio 

(Breznock & Strack, 1982; D. C. Brodbelt, Pfeiffer, Young, & Wood, 2007; Tang, Wu, 

& Peng, 2011). In the absence of evidence-based perioperative fluid rates for animals, the 

guidelines suggest initially starting at 3 mL/kg/hour in cats and 5 mL/kg/hour in dogs, 

based on clinical experience. 

 

The primary risk of providing excessive intravenous fluids in healthy patients is of 

vascular overload, which may actually lead to worsened outcomes, including lung water; 

decreased pulmonary function; coagulation deficits; reduced gut motility; reduced tissue 

oxygenation; increased infection rate; and positive fluid balance, with decreases in 

packed cell volume, total protein concentration, and body temperature (Chappell, Jacob, 

Hofmann-Kiefer, Conzen, & Rehm, 2008; Voldby & Brandstrup, 2016). Supranormal 

fluid resuscitation with 10-30 mL/kg/hour of crystalloid to isoflurane-anaesthetised dogs 

did not improve either urine production or oxygen delivery to tissues (calculated from 

cardiac output measured by thermodilution and oxygen content derived from arterial 

blood gas analysis) , as estimated using cardiac output measured by thermodilution, and 

oxygen content derived from arterial blood gas analysis (Muir, Kijtawornrat, Ueyama, 

Radecki, & Hamlin, 2011). 

 

 



Chapter 1 

 13 

2. CARDIAC FUNCTION AND FLUID LOADING 

2.1 The heart functions 

The main function of the heart is to distribute sufficient oxygenated blood to the entire 

body. First, deoxygenated venous blood returns through the vena cavae to the right atrium 

which then pushes the blood into the right ventricle. Second, the right ventricle pumps 

the blood through the pulmonary artery and then the pulmonary capillaries in the lungs 

where gas exchange takes place. Third, the oxygenated blood moves from the lungs 

through the pulmonary veins to the left atrium, which pushes the blood into the left 

ventricle. Forth, the left ventricle pumps oxygenated blood through the aorta to the organs 

and peripheral tissues. The volume of blood that is pumped from the heart on each 

contraction cycle is called the stroke volume (SV). Cardiac output (CO) is defined as the 

volume of blood that the heart pumps to the systemic circulation per minute, and is 

dependent on heart rate (HR) and SV (CO = HR × SV). SV is influenced by preload, 

contractility, and afterload independently.  

 

The preload is the amount of myocardial stretch prior to each contraction, and clinically 

represents the blood volume in the ventricle or the venous return at the end of diastole. 

For example, haemorrhage decreases the preload, leading to reduced myocardial stretch 

and reduced SV, while fluid administration increases the preload, leading to wide 

myocardial stretch and increased SV (as described by the Frank–Starling law of the heart, 

see below).  

 

Contractility is the intrinsic contractile function of the ventricle and can be defined as the 

force generated by the myocardium independent of preload and afterload. Sympathetic 

nervous stimulation and inotropic drugs such as beta-adrenergic agonists increase the 



Chapter 1 

 14 

contractility, while anaesthetic agents, severe acidosis, hypoxia or hypoxaemia decrease 

the contractility of the myocardium. 

 

Finally, afterload is the pressure that the heart must work against to eject blood during 

systole. Clinically, the afterload is equivalent to the amount of vasoconstriction or 

vascular impedance against which the ventricle ejects. Vasoconstrictors such as alpha-

adrenergic agonists increase the afterload, tending to reduce the SV, while vasodilators 

reduce the afterload, tending to increase the SV.  

 

2.2 The Frank–Starling law of the heart 

Under most conditions, CO is determined by the venous return, which is the rate of blood 

flow into the heart through the vena cavae. A century ago, Dr. Otto Frank and Ernest 

Starling demonstrated increased ventricular contraction when the ventricle was stretched 

prior to contraction. Thus, an increase in ventricular filling pressure, which can be 

achieved by increasing venous return, augmented SV and CO. This is now called the 

Frank–Starling law of the heart. The well-known Frank-Starling curve depicts changes in 

SV or CO in response to changes in venous return, preload, or right atrial pressure (Figure 

1.2-a). An increase in venous return increases the right ventricular filling pressure. 

Increased right ventricular filling pressure stretches sarcomere length in myocytes and 

increasing the sarcomere length increases troponin C calcium sensitivity, which increases 

the rate of cross-bridge attachment and detachment, and the amount of tension developed 

by the muscle fibre, and then augments contraction because the actin and myosin 

filaments are brought to a more nearly optimal degree of overlap for force generation, 

and this subsequently increases SV and thus CO (Guyton & Hall, 2015). Moreover, recent 

muscle physiology research identified a third filament system composed of the giant 



Chapter 1 

 15 

elastic protein titin, which is responsible for most passive stiffness in the physiological 

sarcomere length range. A significant coupling of active and passive forces in cardiac 

muscle, where titin-based passive force promotes cross-bridge recruitment, resulted in 

greater active force production in response to stretch. This focus has been placed on the 

troponin-based “on-off” switching of the thin filament state in the regulation of length-

dependent activation (Fukuda, Terui, Ohtsuki, Ishiwata, & Kurihara, 2009). The Frank- 

Starling curve is also influenced by contractility and afterload, and therefore reflects the 

ventricular function. Figure 1.2-b depicts the curve when the contractility is increased or 

the afterload is reduced, while the curve of Figure 1.2-c depicts the curve when the 

contractility is decreased, or the afterload is increased. If contractility and afterload 

remain constant, the preload determines SV and CO. Thus, when fluid loading increases 

venous return and preload, it will result in an increased SV and CO. Indeed, SV and CO 

cannot increase without an increase in venous return; clarified by the Frank- Starling 

principle that ‘the heart pumps what it receives’. However, excessive preload 

overstretches the myocyte, resulting in unchanged or decreased SV. Although increasing 

preload by fluid loading will increase SV and CO when the patient is on the ascending 

portion of the curve, fluid loading will have little effect when the patient is near the flat 

portion (Guyton, 1955). Thus, if preload is increased when the heart is unable to pump 

the excess volume, it will lead to increased venous pressure and oedema. 

 



Chapter 1 

 16 

 
Figure 1.2 Frank-Starling curve. An increase in preload increase in stroke volume and 

cardiac output. Figure b showed the curve when the contractility is increased or the 

afterload is reduced, while the curve of Figure c is depicted when the contractility is 

decreased, or the afterload is increased. However, the excessive preload overstretches the 

myocyte, resulting in decreased or unchanged stroke volume. 

 

2.3 Fluid responsiveness  

The main purpose of fluid administration is to prevent or treat organ hypoperfusion, 

particularly brain, myocardium, and kidneys, while simultaneously avoiding excessive 

fluid administration. If fluid administration increases SV or CO, then tissue perfusion 

improves, but if the heart is unable to increase SV or CO in response, then fluid loading 

will cause congestion and oedema in the lungs and other tissues. Of critical importance, 

however, is that it is clinically challenging to identify whether tissue hypoperfusion exists, 

or predict if fluid administration will be effective, as subclinical signs of hypoperfusion 

can easily be missed.  Thus, it is typical in clinical practice to administer intravenous 

fluids in ignorance of where the haemodynamic status of a given patient is on the Frank-

Starling curve. Perhaps unsurprisingly then, only roughly half of haemodynamically 

unstable patients respond to a fluid administration, defined as an increase in SV or CO 



Chapter 1 

 17 

upon fluid loading (Marik, Cavallazzi, Vasu, & Hirani, 2009; Michard & Teboul, 2002).  

 

When tissue hypoperfusion is deemed likely, it is essential to find out a patient’s position 

on the Frank-Starling curve to predict whether an increase in SV and CO is to be expected 

from fluid administration. When a patient is located on the steep portion of the Frank-

Starling curve, it is called a fluid “Responder” because increased  SV and thus CO is 

expected after the fluid loading, while a fluid “Non-responder” is on the flat portion of 

the curve because SV and CO will remain unchanged following fluid administration 

(Figure 1.3). Fluid administration in a non-responder will cause a rise in cardiac filling 

pressures and thus hydrostatic pressures in pulmonary and venous circulation causing 

pulmonary and systemic oedema. Thus, a perioperative positive fluid balance was 

associated with increased postoperative mortality or acute kidney injury in intensive care 

unit patients (Oh, Song, Do, Jheon, & Lim, 2019). Conversely, if fluid administration is 

restricted in a patient that would be responsive, it is more likely that the unresolved 

hypotension will be treated with vasopressor therapy. In those under-treated potential 

responders, the vasopressor therapy may impair critical organ perfusion, threaten local 

tissues oxygenation, and cause cardiac arrhythmias which will further reduce CO (Fischer 

et al., 2013; Plurad et al., 2011). This is especially important since supranormal CO and 

oxygen delivery achieved using only  an inotropic drugs  failed to improve postoperative 

outcomes in critically ill patients (Hayes et al., 1994). Therefore, methods for predicting 

fluid responsiveness are required in order to identify those patients who will benefit from 

fluid loading and avoid ineffective and potentially detrimental fluid administration to 

non-responders where positive inotropic drugs may more effective and safer.  



Chapter 1 

 18 

 

Figure 1.3 Fluid responder and non-responder. On the steep portion of the curve, changes 

in preload will result in large changes in stroke volume. A patient on the steep portion of 

the curve would have an increased stroke volume in response to fluid loading. Thus, this 

patient is a fluid responder. On the plateau portion of the curve, identical changes in 

preload will not alter stroke volume as much as they would for a patient on the steep 

portion of the curve. Hence, this patient is a fluid non-responder. 

 

 

 

 

 

 

 

 

 

 

 

 



Chapter 1 

 19 

3. ASSESSMENT OF FLUID RESPONSIVENESS 

3.1 Static predictors of fluid responsiveness 

As mentioned, recent evidence has revealed that outcomes such as pneumonia, acute 

kidney injury, wound infection and hospital length of stay, are improved if fluid therapy 

is individualised based on the patient’s fluid responsiveness (Chong et al., 2018). Medical 

history, clinical examination findings such as skin turgor, blood pressure, pulse rate, and 

urine output, and routine laboratory tests, are important but of limited sensitivity and 

specificity to predict fluid responsiveness (Michard & Teboul, 2002; Vincent & Weil, 

2006). Central venous pressure (CVP) has traditionally been used as an indirect measure 

of venous return or preload in critically ill patients (De Backer & Vincent, 2018).  It was 

believed that CVP is always proportional to preload, and thus a patient with low CVP 

will be a fluid responder, and a patient with high CVP will be a fluid non-responder. 

However, many studies have shown that neither static filling pressures nor the rate of 

change of  static pressures following fluid administration was accurate in predicting fluid 

responsiveness (Magder, 2010; Marik, Baram, & Vahid, 2008). Filling pressures, such as 

CVP, are intramural pressures, and are dependent on atrial and ventricular compliance, 

which varies widely between patients, especially when critically ill (Magder, 2015), and 

are influenced by venous flow. Therefore, CVP is unable to function as a reliable indicator, 

neither of preload nor of fluid responsiveness (Marik et al., 2008). Thus, there is interest 

in developing more reliable predictors of fluid responsiveness. 

 

3.2 Dynamic predictors of fluid responsiveness 

3.2.1 Stroke volume variation 

Over the last decade, dynamic parameters of fluid responsiveness have been described in 

anaesthetised patients that use stroke volume variations (SVV) induced by mechanical 



Chapter 1 

 20 

ventilation. Positive pressure mechanical ventilation transiently compresses the vena 

cavae in the chest, which decreases the venous flow and preload, which in turn reduces 

SV (Mutz et al., 1984; Robotham et al., 1983). The degree to which SV decreases during 

positive pressure ventilation is proportional to the slope of the Frank-Starling curve, in 

the same way that fluid administration may or may not increase SV (Figure 1.4). On the 

steep portion of the curve, a given reduction in preload during insufflation will result in a 

larger SVV than in a patient on the shallower portion of the curve.  Thus, the SVV can be 

used to predict if the patient is a fluid responder.  

 

 

 

Figure 1.4 Stroke volume variation (SVV) A patient on the steep portion of the curve 

would have an increased SV in response to fluid loading. Thus, the patient with large 

SVV is a fluid responder. On the plateau portion of the curve, same changes in preload 

will not alter SV as much as they would for a patient on the steep portion of the curve. 

Hence, the patient with small SVV is a fluid non-responder. 

 

 

In figure 1.5, patients A and B have different curves because of different contractility and 



Chapter 1 

 21 

afterload. Given the same change in absolute preload, patient A has a greater change in 

SV (and greater SVV) than does patient B. So, patient A is a responder even though 

absolute changes in preload are identical for these two patients.  

 

 

 

Figure 1.5 Patients A and B have different curves because of different contractility and 

afterload. Given a same change in absolute preload, patient A has a greater change in SV 

than does patient B. Patient A is a responder and therefore has a greater variation in SV 

even though absolute changes in preload are identical for these two patients.  

 

 

 

To define the SVV of a patient, it is necessary to continuously measure SV or CO by 

echocardiography, or pulse contour analysis (Feissel et al., 2001; Yi, Liu, Qiao, Wan, & 

Mu, 2017) (see Section 4. MEASUREMENT OF CARDIA OUTPUT IN DOGS in 

Chapter 1). However, in veterinary clinical practice, it is challenging to accurately 

measure SV. Therefore, an alternative dynamic parameter reflecting SVV has been 



Chapter 1 

 22 

introduced, called pulse pressure variation (PPV). PPV is obtained directly from the 

peripheral arterial pressure waveform (Yang & Du, 2014). Unlike continuous 

echocardiography, it is feasible to place an arterial line in dogs to measure direct arterial 

pressure during anaesthesia. 

 

3.2.2 Pulse pressure variation (PPV)  

Pulse pressure (PP: difference between systolic and diastolic arterial pressure) is almost 

proportional to SV, and inversely proportional to arterial compliance (Bighamian & Hahn, 

2014). During positive pressure ventilation, arterial compliance remain relatively 

constant over the course of a single breath, thus PP can be used as surrogate of SV, and 

thus PPV is almost identical to SVV. In other words, SVV or PPV induced by positive 

pressure mechanical ventilation will be greater when the patient sits on the steep portion 

of the curve, predicting an increase in SV following fluid loading. PPV is calculated as 

the maximal pulse pressure less the minimum pulse pressure divided by the average of 

these two pressures (Michard et al., 1999)(Figure 5).  

 

!!" = 	100	X	 (Maximum	PP	 − 	Minimum	PP)Average	PP  

 

Where Average PP = (Maximum PP + Minimum PP)/ 2 

 

 

 

 



Chapter 1 

 23 

Figure 1.6 Pulse pressure variation (PPV) PPV, 100 × (PPMax - PPMin)/Average PP; 

Average PP, (PPMax + PPMin)/2; PA, arterial pressure; PAW, airway pressure; PPMax, 

maximum pulse pressure after a positive pressure breath; PPMin, minimum pulse 

pressure after a positive pressure breath (from Scott et al. 2001).  

 

PPV has been shown to be highly predictive of fluid responsiveness in a systematic 

review of its use in people (Marik et al., 2009). The major finding of this systematic 

review and meta-analysis was that PPV during positive pressure mechanical ventilation 

with a tidal volume > 8 mL/kg was an accurate predictor of fluid responsiveness in 

critically ill people, with a sensitivity of 0.88 [95% confidence interval (95%CI): 0.81 to 

0.92)], a specificity of 0.89 (95%CI: 0.84 to 0.92), and summary receiver operating 

characteristic (ROC) area under the curve (AUC) of 0.94 (95%CI: 0.91 to 0.95) (Yang & 

Du, 2014). Another meta-analysis evaluated 22 studies that included a total of 807 

mechanically ventilated patients, to evaluate the value of PPV in predicting fluid 

responsiveness (Hong et al., 2014). This reported a summary ROC AUC for PPV of 0.94 

(95%CI: 0.91 to 0.95), with a pooled sensitivity of 0.88 (95%CI: 0.81 to 0.92) and a 

pooled specificity of 0.89 (95%CI: 0.84 to 0.92) (Yang & Du, 2014). Overall, the median 



Chapter 1 

 24 

cutoff value for PPV (expressed as a percentage) to predict fluid responsiveness in this 

meta-analysis was 12% (interquartile range 10 to 13%) with a sensitivity of 0.88 and a 

specificity of 0.89. The grey zone (below which patients are unlikely to respond, and 

above which they are expected to respond) was identified as being between 10 and 13%. 

 

3.2.3 Pulse pressure variation (PPV) for dogs 

In mechanically ventilated anaesthetised experimental dogs, I found that the PPV 

predicted fluid responsiveness with ROC AUC of 0.85 (95%CI: 0.70 to 1.00, p = 0.038) 

and cutoff values of 11% (sensitivity 79%; specificity 80%) (Chapter 2)(Sano, Seo, et 

al., 2018). At the time of submission of that study for publication, there were no 

previously published studies in dogs.  Subsequently, Fantoni et al (2017) reported the 

predictive value of PPV in 33 and 39 mechanically ventilated anaesthetised dogs that 

underwent orthopaedic surgery and ovariohysterectomy respectively. Those authors 

found the PPV predicted fluid responsiveness with ROC AUC of 0.89 and 0.98, and the 

best cutoff values were deemed to be 15% and 16% (Fantoni et al., 2017). Both the results 

of Fantoni et al and those presented in Chapter 2 are similar to the data in people. A further 

study in 63 client-owned dogs found that a PPV ≥ 13% reliably predicted fluid 

responsiveness in 82.9% of cases (Drozdzynska, Chang, Stanzani, & Pelligand, 2018). 

Therefore, PPV is recommended as a useful clinical tool to detect occult hypovolaemia 

or to prevent hypervolaemia and predict the cardiovascular response to fluid challenge in 

anaesthetised dogs being mechanically ventilated. However, PPV requires invasive 

arterial catheter placement, which may be challenging in small and/or hypotensive dogs.  

Therefore, a non-invasive technique would be desirable in veterinary practices. 

 

 



Chapter 1 

 25 

3.2.4 Pleth Variability Index (PVI) 

PVI is a dynamic parameter that is measured non-invasively, that allows for continuous 

and automated calculation of mechanical ventilation-induced variations in the pulse 

oximetry waveform amplitude (Desebbe et al., 2010). In contrast to measurement of PPV, 

PVI can be calculated from the non-invasive placement of a pulse oximeter probe, which 

provides sufficient information to predict fluid responsiveness. 

 

The perfusion index (PI) is an assessment of the pulsatile strength at the monitoring site. 

The PI is an indirect and non-invasive measure of peripheral perfusion. It is calculated by 

means of pulse oximetry by expressing the pulsatile signal (during arterial inflow) as a 

percentage of the non-pulsatile signal, both of which are derived from the amount of 

infrared light absorbed (Goldman, Petterson, Kopotic, & Barker, 2000; Sun & Huang, 

2014). It is comparable to pulse pressure. 

!8 = 	100	X	 direct	current
alternate	current 

 

PVI can be automatically calculated by anaesthetic monitoring equipment using 

maximum and minimum values of PI during mechanically ventilatory cycles (Sun & 

Huang, 2014) (Figure 6). 

!!" = 	100	X	 (Maximum	PI	 − 	Minimum	PI)Maximum	PI  

 

PVI can predict fluid responsiveness in mechanically ventilated people (Cannesson, 

Delannoy, et al., 2008; Cannesson, Desebbe, et al., 2008). A systematic review and meta-

analysis of 18 studies involving 665 people showed that the PVI has a reasonable ability 

to predict fluid responsiveness with the pooled ROC AUC of 0.88 (95%CI: 0.84 to 0.91), 

the pooled sensitivity of 0.73 (95%CI: 0.68 to 0.78)  and specificity of 0.82 (95%CI: 0.77 



Chapter 1 

 26 

to 0.86) (Chu, Wang, Sun, & Wang, 2016). Furthermore, PVI measured in sedated 

patients prior to induction, has been reported as a good predictor of hypotension following 

general anaesthesia induction (Tsuchiya, Yamada, & Asada, 2010). Therefore, these 

results could be extrapolated to dogs. 

 

 

Figure 1.7 Pleth Variability Index (PVI) PVI, 100 × (PImax - PImin)/PImax; PImax, Maximum 

perfusion index; PImin, Minimum PI (from Masimo HP; https://www.masimo.com/pvi/) 

 

 

3.2.5 Pleth Variability Index (PVI) for dogs 

In mechanically ventilated isoflurane-anaesthetised dogs after premedication with 

acepromazine, PVI predicted fluid responsiveness with ROC AUC of 0.84 (95%CI: 0.68 

to 1.00, p = 0.043) and cutoff values of 9.3% (sensitivity 0. 86; specificity 0.70) (Chapter 

2) (Sano, Seo, et al., 2018). At the time of publication, there were no previous studies in 

dogs, however a subsequent clinical study of 39 mechanically ventilated anaesthetised 

dogs has been published (Celeita-Rodríguez et al., 2019). In that subsequent study, the 

PVI could reliably predict fluid responsiveness with ROC AUC of 0.91 using a cut-off 



Chapter 1 

 27 

value of 11% (Celeita-Rodríguez et al., 2019). PVI can be more easily used in routine 

veterinary practice than PPV, since PVI can be measured noninvasively and does not 

require an arterial catheter, however mechanical ventilation is still needed. Unfortunately, 

availability of mechanical ventilation in veterinary practice is not common (Sano, Barker, 

et al., 2018) and therefore, alternative methods need to be investigated. 

 

3.3 Limitation of dynamic predictors of fluid responsiveness 

The dynamic predictors described above are less reliable to predict fluid responsiveness 

in spontaneously breathing patients than in mechanically ventilated patients (Soubrier et 

al., 2007). However, as stated, mechanical ventilation is not widely available in veterinary 

practice, and also those techniques cannot be used in unanaesthetised animals. Moreover, 

dynamic indices are affected by not only the preload but also other factors such as heart 

rate (Morgan, Abel, Mullins, & Guntheroth, 1966), respiratory rate  (De Backer, Taccone, 

Holsten, Ibrahimi, & Vincent, 2009), pleural pressure (Liu et al., 2016), and tidal volume 

(Díaz, Erranz, Donoso, Salomon, & Cruces, 2015; Kim & Pinsky, 2008). These 

confounding factors may decrease clinical application (Marik & Lemson, 2014). 

Therefore, further investigations to determine their accuracy for the assessment of fluid 

responsiveness are essential. 

 

3.4 Mini-fluid challenge 

The mini-fluid challenge is a strategy to assess fluid responsiveness based on a change in 

SV after a small loading dose of fluid. Theoretically, the change in SV at the steep portion 

of the Frank-Starling curve will be greater than at the plateau portion after the 

administration of a small fluid load, or a larger fluid volume (Figure 1.8). Muller et al 

(2011) showed that there was a good correlation between the increase in SV after 100 mL 



Chapter 1 

 28 

of the mini-fluid challenge, and the increase in SV after 500 mL of the fluid challenge (r 

= 0.81, 95%CI: 0.66 – 0.90 (Muller et al., 2011). Therefore, the magnitude of the change 

in SV after a small fluid loading could predict responsiveness to a larger fluid bolus. 

Studies in people have shown that a mini-fluid challenge could predict fluid 

responsiveness in both mechanically ventilated and spontaneously breathing people 

(Guinot et al., 2015; Muller et al., 2011). Messina et al (2019) published a systematic 

review and a metanalysis of 21 studies of the reliability of the mini-fluid challenge in 

predicting fluid responsiveness in a total of 805 human patients in the intensive care unit 

and operating room. The pooled ROC AUC for the mini-fluid challenge was 0.91 

(95%CI: 0.85 to 0.97) and the pooled sensitivity and specificity were 0.82 (95%CI: 0.76 

to 0.88) and 0.83 (95%CI: 0.77 to 0.89) respectively, with a best threshold of 5% (grey 

zone: 3.0 to 7.0%) (Messina et al., 2019). To my knowledge, the study described in 

Chapter 4 of this thesis is the only one published that has evaluated the mini-fluid 

challenge in dogs (Sano, Fujiyama, et al., 2019), (Sano, Fujiyama, et al., 2019).  In that 

study, I found that the mini-fluid challenge of 3mL/kg was a reliable predictor of fluid 

responsiveness in mechanically ventilated dogs, with a ROC AUC of 0.93 (95%CI: 0.79 

to 1.00), sensitivity of 1.00 and specificity of 0.90. However, that study was in clinically 

healthy dogs, and an assessment of its accuracy in clinical patients will be required before 

widespread use in practice can be recommended.  

 

 



Chapter 1 

 29 

Figure 1.8 The Frank-Starling curve of the heart. Mini-fluid challenge is a strategy to 

assess fluid responsiveness based on a change in stroke volume (SV) after a small loading 

dose of fluid. The change in SV at the steep portion of the curve (responders) will be 

greater than at the plateau portion (non-responders) after both the mini-fluid challenge 

and fluid challenge. Therefore, the magnitude of the change in SV after the mini-fluid 

challenge could predict responsiveness to the fluid challenge. 

 

 

 

 

 

 

 

 

 

 

 

 



Chapter 1 

 30 

4. MEASUREMENT OF CARDIA OUTPUT IN DOGS 

4.1 Clinical gold standard technique 

Measurement of SV and CO can facilitate cardiovascular management during 

anaesthesia. Assessment of trends in SV and CO are also important when treating with 

inotropic, vasoactive drugs and intravenous fluids (Hasanin, 2015). The thermodilution 

technique with a pulmonary artery catheter is considered the clinical standard method to 

measure CO in dogs based on the Stewart–Hamilton equation. It was described (Fegler, 

1954) and tested in dogs using regression analysis, resulting in a good agreement with the 

Fick method (Hendriks, Schipperheyn, & Quanjer, 1978). However, the placement of a 

pulmonary artery catheter is clinically invasive and challenging in veterinary practice and 

associated with increased postoperative complications in human medicine (Harvey et al., 

2005; Sakka, Reinhart, Wegscheider, & Meier-Hellmann, 2000). Therefore, there is a 

need for more clinically feasible yet accurate means of measuring SV and CO in 

veterinary practice. 

 

4.2 Minimally invasive technique 

Several minimally invasive methods for CO measurement have been tested in dogs. 

Arterial pressure contour analysis is a minimally invasive technique for continuous CO 

determination. Systems for the arterial pulse contour analysis require calibration by an 

invasive technique such as a thermodilution technique, which has been performed for the 

PiCCO (Pulsion Medical Systems, Germany) (Garofalo et al., 2016; Hofer, Cecconi, 

Marx, & della Rocca, 2009) and the LiDCO (LiDCO Ltd., Cambridge, UK)(Mathews & 

Singh, 2008) in people. However, comparisons between those systems and with the gold 

standard thermodilution technique in dogs have produced conflicting results (Garofalo et 

al., 2016; Morgaz et al., 2014). The FloTrac/Vigileo system (Edwards Lifesciences AG, 



Chapter 1 

 31 

Switzerland), without prior calibration, uses the arterial pressure contour to continuously 

monitor and/or calculate SV, CO, SVV and systemic vascular resistance through a 

standard peripheral arterial line based on algorithm using the relationship between aortic 

pulse pressure, SV and aortic compliance (Bektas et al., 2012). Although this system is 

less invasive, it requires an arterial line and did not have a good agreement with the 

thermodilution technique in dogs (Bektas et al., 2012). 

 

Ultrasonography is able to estimate blood flow through a valve orifice using colour-flow 

Doppler. The area under the velocity-time curve can be calculated automatically in an 

ultrasound machine. This area under the curve is called the velocity time integral, and it 

is proportional to how far blood moves during the time period. If the valve orifice is 

assumed to be circular, the orifice area can be calculated using the diameter (area = π × 

radius2), which can be measured using ultrasonography. The formula for SV is area 

multiplied by VTI (SV = area × VTI). CO can be calculated by heart rate multiplied by 

SV (CO = HR × VTI). Importantly, ultrasonography is non-invasive, and transthoracic 

echocardiography at the main pulmonary artery provides an accurate measurement of CO 

with an excellent correlation with thermodilution techniques in dogs (Lopes et al., 2010). 

Similarly, transoesophageal echocardiography provides good agreement with the 

thermodilution technique in dogs (Mantovani et al., 2017; Yamashita et al., 2007) and 

transoesophageal Doppler devices accurately reflected the direction and magnitude of the 

changes of CO over time during abrupt hemodynamic changes in dogs (de Figueiredo, 

Cruz, Silva, & Rocha, 2004). However, all of these techniques require an expensive 

machine and an expert with a high level of skill and are thus not easily applicable to 

general veterinary practice. 

 



Chapter 1 

 32 

4.3 Pulse wave transit time 

Pulse wave transit time (PWTT) is the time from the electrocardiogram (ECG) R-wave 

peak to the rise point of the pulse oximeter wave (Sugo et al., 2010). The rise point of the 

pulse wave is defined as the point at which the differentiated pulse wave reached 30% of 

its peak amplitude (Figure 1.9). PWTT, a measure of velocity,  has proven to be inversely 

proportional to SV and has a strong correlation with SV in dogs (Sugo et al., 2010) and 

people (Ishihara et al., 2004). Based on this relationship, Sugo et al. developed a system 

to estimate SV and CO using PWTT (Sugo, Sakai, Terao, Ukawa, & Ochiai, 2012). This 

requires an initial 3-minute period of stable haemodynamics for calibration against 

another CO measurement system, or an automatic patient information calibration based 

on the patient’s information, and their arterial pulse pressure (Ishihara et al., 2004), which 

is similar to arterial pressure contour analysis. It is easy to use, inexpensive, minimally 

invasive and requires only routine anaesthetic monitoring available in most veterinary 

clinics (pulse oximetry, ECG, non-invasive or invasive arterial blood pressure 

monitoring). In a canine experiment, the correlation of estimated CO with CO obtained 

using electromagnetic flow meters on the aorta, was high (r = 0.825), however the 

agreement between the two methods was not clearly reported (Sugo et al., 2010). Because 

a high correlation does not always indicate good agreement between the two methods, 

precision of agreement with percentage error using Bland–Altman analysis should be 

used to assess interchangeability of two methods (Critchley, Lee, & Ho, 2010).  

 



Chapter 1 

 33 

 

Figure 1.9 Pulse Wave Transit Time (PWTT). PWTT was calculated as the time from 

the ECG R-wave peak to the rise point of the pulse oximeter wave. The rise point of the 

pulse wave was defined as the point at which the pulse wave reached 30% of its peak 

amplitude. 

 

 

 

 

 

 

 

 

 

 

 

 

 



Chapter 1 

 34 

5. PERIOPERATIVE HAEMODYNAMIC MANAGEMENT IN DOGS 

5.1 Purpose of perioperative haemodynamic management 

Anaesthesia is essential when animals need major or minor surgeries, endoscopy, 

diagnostic imaging, or other invasive diagnostic techniques. Anaesthesia provides 

unconsciousness, amnesia, muscle relaxation, and analgesia to animals, allowing these 

procedures to be performed humanely without pain or excessive physiological responses. 

However, haemodynamic depression can be caused concurrently, as anaesthetic agents 

are usually cardiovascular depressants (D. Brodbelt, 2009; Gaynor et al., 1999). 

Haemodynamic depression may preclude sufficient oxygen delivery (DO2) to tissues, 

causing tissue hypoxia. Inappropriate perioperative haemodynamic management has been 

shown to cause severe tissue hypoxia during anaesthesia, resulting in many clinical 

complications such as surgical infection, delayed wound healing, prolonged 

hospitalization, and multiple organ dysfunction syndrome in dogs (Snowdon, Smeak, & 

Chiang, 2016; Turk, Singh, & Weese, 2015) and people (Monk, Saini, Weldon, & Sigl, 

2005; Tassoudis et al., 2011).  

 

The main goal of perioperative haemodynamic management is to provide adequate DO2 

to major organs such as the brain, heart, and kidneys, and to peripheral tissues. DO2 

depends on haemoglobin (Hb) concentration, saturation of arterial Hb with oxygen 

(SaO2), and CO. CO is directly reduced by anaesthetic agents, but Hb and SaO2 are not. 

Assuming normal Hb and SaO2 (although both can be reduced under anaesthesia under 

some circumstances), CO determines DO2. Therefore, optimisation of CO will lead to 

successful perioperative haemodynamic management. Consistent with that, a systematic 

review and meta-analysis showed that optimisation of CO has been shown to decrease 



Chapter 1 

 35 

perioperative morbidity and hospital length of stay in cardiac patientss (Aya, Cecconi, 

Hamilton, & Rhodes, 2013). 

 

5.2 Perioperative blood pressure management 

As described above, it is challenging to measure CO in veterinary practices. Thus, instead 

of monitoring CO, blood pressure, especially mean arterial pressure (MAP), is used as a 

surrogate measure of tissue perfusion in dogs and people (Monk et al., 2005). Cerebral 

blood flow autoregulation is supposed to maintain normal perfusion between a MAP of 

60 and 150 mm Hg (Figure 1.10) (Dagal & Lam, 2009; Paulson, Strandgaard, & 

Edvinsson, 1990) whilst autoregulation of renal perfusion flow drops off steeply at a MAP 

of 70 mmHg (Shipley & Study, 1951) (Figure 1.11) . Thus, MAP needs to be maintained 

at least 60 mmHg to prevent brain ischaemia, or more than 70 mmHg to provide enough 

blood perfusion to maintain normal kidney function during anaesthesia (Iizuka, Kamata, 

Yanagawa, & Nishimura, 2013; Redondo et al., 2007). However, most anaesthetic agents 

have vasodilatory and negative inotropic effects that can cause hypotension (D. Brodbelt, 

2009; Gaynor et al., 1999). In fact, severe hypotension (MAP < 60 mmHg or systolic 

arterial pressure < 80 mmHg) occurs in up to 65% of anaesthetised in dogs and cats 

(Iizuka et al., 2013; Redondo et al., 2007). Therefore, monitoring of MAP and the prompt 

treatment of low MAP is necessary. 

 

 

 



Chapter 1 

 36 

 
Figure 1.10 Cerebral blood flow autoregulation (Paulson et al., 1990). It typically 

operates between MAP of the order of 60 and 150 mmHg (normal MAP). 

 

 

Figure 1.11 Renal perfusion flow autoregulation (Shipley & Study, 1951). Renal 

perfusion flow (RPF) drops off at 70 mmHg but the plateau really starts at about 110 

mmHg. Glomerular filtration (GFR) plateaus at 120 mmHg and by the time renal artery 

mean blood pressure is 70 mmHg the GFR is at about 70%. 

 

5.3 Treatment for hypotension in small animals 



Chapter 1 

 37 

Hypotension under anaesthesia is a frequent occurrence, even in healthy patients. When 

observed, assessment of anaesthetic depth should be the first action because deep 

anaesthesia is a common cause (Monk et al., 2005). Probably the most common treatment 

of hypotension if anaesthetic depth is deemed appropriate is the administration of a fluid 

bolus. However, caution should be taken when using fluid therapy as the sole method to 

correct anaesthesia-related hypotension as high rates of fluids can exacerbate 

complications rather than prevent them (Muir et al., 2011; Voldby & Brandstrup, 2016). 

Based on 2013 AAHA/ AAFP fluid therapy guidelines in dogs and cats  (Davis et al., 

2013), the following process is the recommended treatment of hypotension in the 

anaesthetised dogs:  

1. Decrease anaesthetic depth and/or inhalant concentration.  

2. Provide an intravenous bolus of an isotonic crystalloid such as lactated Ringer’s 

solution (3–10 mL/kg). Repeat once if needed.  

3. If response is inadequate, consider intravenous administration of a colloid such as 

hetastarch. Slowly administer 5–10 mL/kg for dogs and 1–5 mL/kg for cats, 

titrating to effect to minimize the risk of vascular overload (measure blood 

pressure every 3–5 minutes)  (Chappell et al., 2008).  Colloids are more likely to 

increase blood pressure than crystalloids (Aarnes, Bednarski, Lerche, Hubbell, & 

Muir, 2009). 

4. If response to crystalloid and/or colloid boluses is inadequate and patient is not 

hypovolemic, techniques other than fluid therapy may be needed (e.g., inopressors 

or, balanced anaesthetic techniques) (Chappell et al., 2008). 

As described above, excessive fluid administration can result from the treatment of 

hypotension in the anaesthetised dogs. Therefore, prevention of hypotension using 

inopressors such as noradrenaline could reduce the incidence of fluid overload in clinical 



Chapter 1 

 38 

veterinary practices (Sano, Chambers, & Bridges, 2019). However, peripheral perfusion 

may be impeded due to excessive vasoconstriction, which can be severe enough to cause 

peripheral gangrene (Kwon, Hong, & Park, 2018). 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



Chapter 1 

 39 

6 AIM AND OBJECTIVES OF THESIS 

This literature review established that although fluid therapy is important in order to 

improve haemodynamics in dogs and people, excessive fluid administration is 

detrimental. The clinical challenge is to determine which patient will respond to fluid 

administration, and to determine the correct amount of fluid for each patient. Currently, 

there are several methods to determine fluid responsiveness in people but studies in dogs 

are scarce. Dynamic parameters to predict fluid responsiveness are often used in people 

but have many limitations. Application of dynamic parameters to dogs is possible but 

clinically limited because of their invasiveness and the requirement for a positive pressure 

mechanical ventilator. The mini-fluid challenge is an alternative method to identify fluid 

responsiveness and may be used in spontaneously breathing dogs. However, reliable 

measurement of the change in stroke volume is necessary. Pulse wave transit time, which 

can be clinically obtained non-invasively in dogs, may be used as a surrogate parameter 

for SV. The mini-fluid challenge using pulse wave transit time has the potential to detect 

fluid responsiveness in dogs non-invasively. However, those methods are still 

cumbersome in clinical veterinary practice because there is currently a lack of suitable 

equipment. Simple prophylactic noradrenaline infusion may be able to prevent 

hypotension in anaesthetised dogs and reduce the incidence of fluid overload caused by 

the typical treatment of hypotension in veterinary practice: a bolus of fluid.  

 

6.1 Thesis aim 

The aim of the research presented in this thesis was to develop clinically feasible methods 

to determine fluid responsiveness and prevent fluid overload in anaesthetised dogs.  

 

6.2 Thesis Objectives 



Chapter 1 

 40 

The investigation of clinical methods to prevent excessive fluid administration was 

described by the following objectives: 

Chapter 2 To evaluate whether PPV and PVI are more accurate than CVP for 

predicting fluid responsiveness in mechanically ventilated isoflurane-

anaesthetised dogs after premedication with acepromazine. 

Chapter 3 To evaluate the ability of PWTT to detect changes in SV and to estimate 

CO compared with the thermodilution technique in isoflurane-

anaesthetised dogs. 

Chapter 4 To investigate whether percentage changes in PWTT induced by mini-

fluid challenges predict fluid responsiveness in mechanically ventilated 

anaesthetised dogs. 

Chapter 5 To investigate whether percentage changes in PWTT following mini-fluid 

challenge could predict fluid responsiveness in spontaneously breathing 

anaesthetised dogs. 

Chapter 6 To investigate whether noradrenaline infusion prior to hypotension 

improves anaesthetic management in dogs undergoing 

ovariohysterectomy. 

Three experimental studies in healthy dogs, and two clinical trials in client owned healthy 

dogs were conducted to achieve these objectives. 

 

 

 

 

 

 



Chapter 1 

 41 

Abbreviations 

AUC  area under a curve 

CO  cardiac output (mL) 

CVP  central venous pressure (mmHg) 

DO2  oxygen delivery (mL/minute) 

Hb  haemoglobin (g/dL) 

HR  heart rate (beat/minute) 

MAP  mean arterial pressure (mmHg) 

PI  perfusion index 

PPV  pulse pressure variation (%) 

PVI  pleth Variability Index (%) 

PWTT  pulse wave transit time (msecond) 

ROC  receiver operating characteristic 

SV  stroke volume (mL) 

SVV  stroke volume variation (%) 

VTI  velocity time integral (cm) 

95%CI  95% confidence interval 

 

 

 

 

 

 

 

 



Chapter 1 

 42 

7 REFERENCES  

Aarnes, T. K., Bednarski, R. M., Lerche, P., Hubbell, J. A., & Muir, W. W., 3rd. (2009). 
Effect of intravenous administration of lactated Ringer's solution or hetastarch for 
the treatment of isoflurane-induced hypotension in dogs. Am J Vet Res, 70(11), 
1345-1353. doi:10.2460/ajvr.70.11.1345 

Arieff, A. I. (1999). Fatal postoperative pulmonary edema: pathogenesis and literature 
review. Chest, 115(5), 1371-1377. doi:10.1378/chest.115.5.1371 

Aya, H. D., Cecconi, M., Hamilton, M., & Rhodes, A. (2013). Goal-directed therapy in 
cardiac surgery: a systematic review and meta-analysis. Br J Anaesth, 110(4), 
510-517. doi:10.1093/bja/aet020 

Bednarski, R., Grimm, K., Harvey, R., Lukasik, V. M., Penn, W. S., Sargent, B., & Spelts, 
K. (2011). AAHA anesthesia guidelines for dogs and cats. J Am Anim Hosp Assoc, 
47(6), 377-385. doi:10.5326/jaaha-ms-5846 

Bektas, R. N., Kutter, A. P., Hartnack, S., Jud, R. S., Schnyder, M., Matos, J. M., & 
Bettschart-Wolfensberger, R. (2012). Evaluation of a minimally invasive non-
calibrated pulse contour cardiac output monitor (FloTrac/Vigileo) in 
anaesthetized dogs. Vet Anaesth Analg, 39(5), 464-471. doi:10.1111/j.1467-
2995.2012.00741.x 

Bighamian, R., & Hahn, J. O. (2014). Relationship between stroke volume and pulse 
pressure during blood volume perturbation: a mathematical analysis. Biomed Res 
Int, 2014, 459269. doi:10.1155/2014/459269 

Bishop, M. H., Shoemaker, W. C., Appel, P. L., Meade, P., Ordog, G. J., Wasserberger, 
J., . . . et al. (1995). Prospective, randomized trial of survivor values of cardiac 
index, oxygen delivery, and oxygen consumption as resuscitation endpoints in 
severe trauma. J Trauma, 38(5), 780-787. doi:10.1097/00005373-199505000-
00018 

Bishop, M. H., Shoemaker, W. C., Appel, P. L., Wo, C. J., Zwick, C., Kram, H. B., . . . 
Fleming, A. W. (1993). Relationship between supranormal circulatory values, 
time delays, and outcome in severely traumatized patients. Crit Care Med, 21(1), 
56-63. doi:10.1097/00003246-199301000-00013 

Boyd, O., Grounds, R. M., & Bennett, E. D. (1993). A randomized clinical trial of the 
effect of deliberate perioperative increase of oxygen delivery on mortality in high-
risk surgical patients. Jama, 270(22), 2699-2707.  

Brandstrup, B., Svensen, C., & Engquist, A. (2006). Hemorrhage and operation cause a 
contraction of the extracellular space needing replacement--evidence and 
implications? A systematic review. Surgery, 139(3), 419-432. 
doi:10.1016/j.surg.2005.07.035 

Brandstrup, B., Tonnesen, H., Beier-Holgersen, R., Hjortso, E., Ording, H., Lindorff-
Larsen, K., . . . Pott, F. (2003). Effects of intravenous fluid restriction on 
postoperative complications: comparison of two perioperative fluid regimens: a 
randomized assessor-blinded multicenter trial. Ann Surg, 238(5), 641-648. 
doi:10.1097/01.sla.0000094387.50865.23 

Breznock, E. M., & Strack, D. (1982). Blood volume of nonsplenectomized and 
splenectomized cats before and after acute hemorrhage. Am J Vet Res, 43(10), 
1811-1814.  

Brodbelt, D. (2009). Perioperative mortality in small animal anaesthesia. Vet J, 182(2), 
152-161. doi:10.1016/j.tvjl.2008.06.011 

Brodbelt, D. C., Pfeiffer, D. U., Young, L. E., & Wood, J. L. (2007). Risk factors for 
anaesthetic-related death in cats: results from the confidential enquiry into 



Chapter 1 

 43 

perioperative small animal fatalities (CEPSAF). Br J Anaesth, 99(5), 617-623. 
doi:10.1093/bja/aem229 

Bundgaard-Nielsen, M., Secher, N. H., & Kehlet, H. (2009). 'Liberal' vs. 'restrictive' 
perioperative fluid therapy--a critical assessment of the evidence. Acta 
Anaesthesiol Scand, 53(7), 843-851. doi:10.1111/j.1399-6576.2009.02029.x 

Cannesson, M., Delannoy, B., Morand, A., Rosamel, P., Attof, Y., Bastien, O., & Lehot, 
J. J. (2008). Does the Pleth variability index indicate the respiratory-induced 
variation in the plethysmogram and arterial pressure waveforms? Anesth Analg, 
106(4), 1189-1194, table of contents. doi:10.1213/ane.0b013e318167ab1f 

Cannesson, M., Desebbe, O., Rosamel, P., Delannoy, B., Robin, J., Bastien, O., & Lehot, 
J. J. (2008). Pleth variability index to monitor the respiratory variations in the 
pulse oximeter plethysmographic waveform amplitude and predict fluid 
responsiveness in the operating theatre. Br J Anaesth, 101(2), 200-206. 
doi:10.1093/bja/aen133 

Celeita-Rodríguez, N., Teixeira-Neto, F. J., Garofalo, N. A., Dalmagro, T. L., Girotto, C. 
H., Oliveira, G. C. V., & Santos, I. F. (2019). Comparison of the diagnostic 
accuracy of dynamic and static preload indexes to predict fluid responsiveness in 
mechanically ventilated, isoflurane anesthetized dogs. Vet Anaesth Analg, 46(3), 
276-288. doi:10.1016/j.vaa.2018.12.004 

Chappell, D., Jacob, M., Hofmann-Kiefer, K., Conzen, P., & Rehm, M. (2008). A rational 
approach to perioperative fluid management. Anesthesiology, 109(4), 723-740. 
doi:10.1097/ALN.0b013e3181863117 

Chong, M. A., Wang, Y., Berbenetz, N. M., & McConachie, I. (2018). Does goal-directed 
haemodynamic and fluid therapy improve peri-operative outcomes?: A systematic 
review and meta-analysis. Eur J Anaesthesiol, 35(7), 469-483. 
doi:10.1097/eja.0000000000000778 

Chu, H., Wang, Y., Sun, Y., & Wang, G. (2016). Accuracy of pleth variability index to 
predict fluid responsiveness in mechanically ventilated patients: a systematic 
review and meta-analysis. J Clin Monit Comput, 30(3), 265-274. 
doi:10.1007/s10877-015-9742-3 

Coller, F. A., Dick, V. S., & Maddock, W. G. (1936). Maintenance of normal water 
exchange with intravenous fluids. Journal of the American Medical Association, 
107, 1522–1527.  

Corcoran, T., Rhodes, J. E., Clarke, S., Myles, P. S., & Ho, K. M. (2012)