EVALUATION OF THE USE OF INDIRECT CALORIMETRY FOR THE
MEASUREMENT OF RESTING ENERGY EXPENDITURE IN DOGS
A Thesis
Presented to
The Faculty of Graduate Studies
of
The University of Guelph
by
ELIZABETH O'TOOLE
In partial fuIfilment of requirements
for the degree of
Doctor of Veterinary Science
August, 2000
O Elizabeth O'Toole, 2000 Bibliothèque nationale du Canada Acquisitions and Acquisitions et Bibliographie Services services bibliographiques 395 Wellington Street 395, rue Wellington Ottawa ON K1A ON4 Ottawa ON KIA ON4 Canada Canada YOMfi& votre référence
Ourlüe Noire rdfbrence
The author has granted a non- L'auteur a accordé une licence non exclusive licence allowing the exclusive permettant à la National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, distribute or sell reproduire, prêter, distribuer ou copies of this thesis in microform, vendre des copies de cette thèse sous paper or electronic formats. la forme de microfiche/nlm, de reproduction sur papier ou sur format électronique.
The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts fiom it Ni la thèse ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation. ABSTRACT
EVALUATION OF THE USE OF INDIRECT CALORIMETRY FOR THE
MEASUREMENT OF RESTING ENERGY EXPENDITURE IN DOGS
Elizabeth O'Toole Advisors: University of Guelph Dr. K.A. Mathews Dr. C. Miller
This thesis investigates the utility of indirect calorimetry to rneasure the resting energy expenditure (REE) in dogs. Until recently the only available method to assess the 24-hour energy expenditure (EE) in animals has been to estimate it with a predictive equation; 70 (BW in kg) O-" corrected for the severity of illness or injury
(IER). Providing an alternative rnethod for assessing the 24-hour EE, which was comparable to a previous clinical standard, a closed-circuit spirometer, would provide veterinarians with a more reliable determination of the REE. The objectives of this study were: 1) to compare the level of agreement between the oxygen consurnption
(V02/kg) as detemined by the indirect calorimeter and the VOdkg as determined by a closed-circuit spirometer in healthy dogs; 2) to confirm the reliability of indirect calorimetry in healthy dogs; 3) to determine the Ievel of agreement for REE via two rnethods, indirect calorirnetry and the predictive equation, in healthy and il1 or injured dogs.
Gas exchange measurements of V02 and carbon dioxide production (VC02), were perforrned by a portable indirect calorimeter. The REE in kilocalories per day was calculated by the application of the abbreviated Weir formula. lnterclass correlation coefficients were used to demonstrate the reliabil'w of the serial rneasurements of V02 and REE in seven healthy staffawned dogs on two consecutive days. Bland-Altman analysis of agreement was used to determine the level of agreement between the indirect calorimeter and the closed-circuit spirometer in six anesthezied healthy research dogs. This method was also utilised to demonstrate the level of agreement between the two methods of REE determination; measured REE and predicted REE in healthy dogs, and in dogs with either medical illness or recovering from major surgery or major trauma.
This study demonstrates that there was a good level of agreement between the indirect calorimeter and the ~Iosed-circuitspirometer for the measurement of
V02/kg in dogs. A reliable measure of Vonand REE was obtained in healthy dogs using open-flow indirect caiorirnetry. There was a poor level of agreement between the measured and predicted REE in each group of il1 or healthy dogs studied. The predictive equation only agreed to within + 20% of the measured REE between 50 to
56% of the time on the two days studied. ACKNOWLEDGEMENTS
I would like to thank my graduate committee rnembers, Drs. Karol Mathews,
Craig Miller, Brian A. Wilson, Tony Abrams-Ogg and Clive Davis for their support and
assistance in this thesis project. 1 would like to thank William Sears for providing
invaluable advice concerning the data analyses and Dr Brenda Bonnett for providing
an explanation of this advice and for placing the results in to a clinical context. I would also like to thank Wayne McDonnell for his support, questions and passion in the quest for scientific knowledge.
Financial support for this research project was provided by the Ontario
Veterinary ColIege Pet Trust Fund. 1 would like to congratulate both this foundation and the Natasha Fellowship for their commitment to research and further graduate training of veterinarians respectively.
Thanks, in no small degree, is also due to Jackie Gordon and the ICU technicians (J Bali, H Creasy, K Taylor, A Steele, D Cloutier, A Archer, and M Brooks) who assisted in the collection of the data. Thank you for al1 your support especially,
Jackie to whom the machines seemed to love. In addition, I would aIso like to thank
Mark Smith, from the Aerosport Inc, for his time, knowladge, encouragement and generosity. Without his support I would have not been able to complete my research project in a tirnely manner.
Finally, I would sincerely like to thank my parents, Eamon and Nicki, well for just being there and for being my parents. 1 need to thank Isabelle, Debbie, Carolyn,
Rosemary, and Tom and Cynthia for being friends. They have provided me with love, support and acceptance, which I am extremely grateful for. 1 also need to thank Joan
Coulter for her advice and objective viewpoints during my three-year residency period. Lastiy, I must acknowledge the O'Toole menagerie that knows how to Iive and how to enjoy the small things in life. DECLARATION OF WORK PERFOMRED
I declare that with the exception of the items outline below, al1 work reported in this
thesis was performed by me.
The (CU technicians and a sumrner student, Jackie Gordon, assisted in the collection
of the measurements of the resting energy expenditure in the il1 and healthy dogs. A
second summer student, Dorothy Kitchen, performed the monitoring during the
anaesthesia of the healthy dogs. Complete blood counts biochemistry profiles, free senim thyroxine levels in the healthy dogs and the CBC, and biochemical profiles in the il1 dogs were performed by the technical staff of the Animal Health Laboratory,
Guelph, Ontario. iii
TABLE OF CONTENTS
ACKNOWLEDGMENTS i
DECLARATION OF WORK PERFORMED ii
LIST OF TABLES viii
LIST OF FIGURES ix
Chapter 1: Review of the Literature 1
1.O INTRODUCTION 1
1.1 GOALS AND HYPOTHESIS 1
1.2 LITERATURE REVIEW 2
1-2.1 Introduction 2
1.2.2 Starvation Metabolism 3
1.2.3 Stress Hypermetabolism - 6
1.2.4 Malnutrition 9
1.2.5 Ovetfeeding 11
1.2.6 Daily Energy Expenditure 11
1.2.7 Estimated Energy Expenditure 18
1.2.8 Measured Energy Expenditure 23
1.2.9 Theoretical Considerations of Indirect Calorirnetry 25
Indirect Calorimetry Equations 27
HaIdane Transformation 27
1.2.1 0 Clinical Application of Indirect Calorimetry 28
1.2.1 1 Summary 32
1-2.1 2 Reference List 33 Chapter 2: Evaluation of the Accuracy and Reliability of lndirect Calon'rnetry for the Measurernent of Resting Energy Expenditure in Healthy Dogs 47
2.0 Abstract 47
2.1 Introduction 49
2.2 Materiai and Methods 50
2.3 Results 55
2.4 Discussion 57
2.5 Reference List 70
2.6 Addendum for Manuscript 1 75
Chapter 3: Cornparison of Indirect Calon'metry and the Predictive Equation for the determination of Resting Energy Expenditure in healthy and il1 dogs.
3.0 Abstract 78
3.1 Introduction 79
3.2 Material and Methods 82
3.3 Results 87
3.4 Discussion 90
3.5 Reference List 103
3.6 Addendum for Manuscript # 2 107
Chapter 4: General Conclusions and Future Studies 113
Chapter 5: Appendices 113
5.1. Appendix I. The operating specifications for open-flow indirect calorimetry, TEEM
100, utilised in this thesis. 114 5.2. Appendix Il. Level of agreement between the two methods of indirect calorimetry, open-flow indirect calorimetry and closed-circuit spirometry for the measurement of oxygen consurnption per kilogram in six healthy dogs. 115
5.3. Appendix Ill. Reliability study; MREE by open-flow indirect calorimetry in seven healthy staff owned dogs. - 116
5.4. Appendix IV. Reliabiiity study; oxygen consumption by open-flow indirect calorimetry in seven healthy staff owned dogs. 217
5.5. Appendix V. Reliability study; variance components associated with the oxygen consum ption measurements in seven healthy staff owned dogs.
5.6. Appendix VI. Reliability study; unadjusted mean oxygen consumption per kilogram by open-flow indirect calorimetry in the seven healthy staff-owned dogs on the two consecutive days. 119
5.7. Appendix VII. Signalment and underlying disease in the 77 dogs enrolled in the agreement study between the two methods of REE determination; MREE and PREE. - 120
5.8. Appendix VIII. Agreement study between MREE and PREE; signalment, gas exchange measurements, MREE and PREE for the il1 and healthy dogs in each group when the readings perfomed in a 24-hr period had been combined.
124
5.9. Appendix IX. Agreement study between MREE and PREE; the level of clinical agreement for the measured REE compared to the predicted REE on an individual dog basis in the four groups on each day. - - 131
5.10. Appendix X. Agreement study between MREE and PREE; MREE and PREE when al1 the indirect calorimetry readings were averaged over the 48-hour period.
139 t 5.1 1. Appendix XI. Agreement study between MREE and PREE; MREE per kilogram
basis in each dog when the readings were averaged over the 48-hour period.
142
5.12. Appendix XII. Agreement study between MREE and PREE; graphical
representation of the cornparison of the MREE and PREE over the 48-hour period in
each group. 143
5.13. Appendix XIII. Agreement study between MREE and PREE; adjusted means of the MREE and PREE for each group over the 48-hour period.
144
5.14. Appendix XIV. Agreement study between MREE and PREE; the level of clinical agreement of al1 groups between the MREE and the PREE when al1 measures were combined over the 48-hour period. 145
5.15. Appendix XV. Agreement study between MREE and PREE; the Ievel of clinical agreement between the MREE and PREE for each group over the 48-hour period.
146
5.1 6. Appendix XVI. Agreement study between MREE and PREE; percentage of data points in each group which lay within and outside the acceptable Iimits of agreement
(+ 20%). 150
5.17. Appendix XVII. Agreement study between MREE and PREE; grand means per group of the oxygen consumption, oxygen consumption per kilogram and respiratory quotient of the 77 dogs on the two days enrolled into the agreement study between the two methods of REE detemination. 151
5.1 8 Appendix XVI I1. Agreement study between MREE and PREE; graphical representation of the grand rneans of the oxygen consumption per kilogram in each group averaged over al1 the readings on the two days. 152 vii
5.19. Appendix XVIX. Agreement study between the two different formulae for REE
calculation; signalment and undedying disease of the 26 dogs with medical disease or
recovering from major surgery or a traumatic incident, which compared the calculation
of the REE including (Weir formula) and excluding protein metabolism (Abbreviated
Weir formula) 153
5.20. Appendix XX. Agreement study between the two different formulae for REE
calculation; the level of agreement between the calculated REE including protein
metabolism compared to the calculated REE excluding protein metabolism.
. - 155
5.21. Appendix XXI. Agreement study between the MREE and PREE and the two different formulae for REE calculation; signalment, group, gas exchange
rneasurements, measured and predicled REE (kcawday), vital signs, blood gas analysis and serurn lactate concentrations, in 77 dogs. This is the non-transformed data. 156 viii
LIST OF TABLES
Table 2.1 : Pearson correlation coefficients and probability values for the pre and post
indirect calorirnetry readings when compared to the spirometer. - 68
Table 2.2: The interclass correlation coefficient for the RE€ of the averaged six-
minute readings in the reliability study over the two days. 69
Table 2.3: The interclass correlation coefficient for the oxygen consumption of the
averaged six-minute readings in the reiiability study over the two days. - 70
Table 2.4: Estimate of the variance components associated with the resting energy
expenditure (REE) measurernents in the reliability study over the two days.
77
Table 3.1 :The level of agreement study between the two deteminations of REE;
measured versus predicted, unadjusted rnean, median, standard deviation and range
of the descriptive statistics; age, weight, illness factor and respiratory quotient on a
per group basis. 97
Table 3.2: The level of agreement study between the two determinations of REE;
measured versus predicted, gender distribution per group of the 77 dogs.
98
Table 3.3: The level of agreement study between the two determinations of REE; measured versus predicted, percentage of data points, which represents the individual dogs REE determinations that are within and outside the clinical acceptable lirnits (+ 20%) on a per group basis. 102
Table 3.4: The adjusted group rneans + 95% confidence interval for the measured
REE on a per kilograrn basis for the 77 dogs eniolled in the clinical study.
1O7
Table 3.5: The adjusted group means + 95% confidence interval for the predicted
REE on a per kilograrn basis for the 77 dogs enrolled in the clinical study,
1O8 LIST OF FIGURES
Figure 1.1 : Theoretical considerations in the appIication of the indirect calorimetry
technique for the calculation of the resting energy expenditure.
Figure 2.1: Agreement between the pre-indirect calorimetry oxygen consumption per
kilogram (V021kg) reading and closed-circuit spirornetry V02/kg reading in six healthy - anaesthetised dogs, 63
Figure 2.2: Agreement between the post-indirect calorimetry oxygen consumption per
kitogram (VOdkg) reading and closed-circuit spirometry V02/kg reading in six healthy
anaesthetised dogs 64
Figure 2.3: Reliability study; mean and the standard error of the mean of resting
energy expenditure in the seven healthy staff- owned dogs. 65
Figure 2.4: Reliability study; mean and the standard error of the mean of oxygen
consumption in the seven healthy staff- owned dogs. 66
Figure 2.5: ReIiability study; difference in the resting energy expenditure, between the
days, over the senal measurements in the seven healthy staff-owned dogs. 67
Figure 2.6: Reliability study; REE of each of the seven healthy staff-owned dogs on the first day. 75
Figure 2-7:Reliability study; REE of each of the seven healthy staff-owned dogs on the second day. 76
Figure 3.1 :The level of agreement study between the two determinations of RE€; measured versus predicted, comparison of the adjusted mean MREE and on a per group basis on each day in the 77 dogs enrolled. 99 Figure 3.2: The level of agreement study between the two deteminations of REE;
measured versus predicted, clinical agreement analysis on the first day for the MREE
compared to the PREE for the 77 dogs on a per group basis. IO0
Figure 3.3: The level of agreement study behiveen the two determinations of REE; measured versus predicted, clinical agreement analysis on the second day for the
MREE compared to the PREE for the 77 dogs on a per group basis. 101 Chanter 1: Review of the Literature
1-0 INTRODUCTION
Malnutrition and overfeeding have been found to adversely affect the clinical outwme
of cntically il1 patients'. Therefore. an accurate assessment of resting energy
- expenditure (REE) is an important wmponent of therapy in human critical care
medicinez3. The predictive equation of REE used in human medicine. even when
corrected for illness and activity factors, differ significantly from the measured RE€. A
more reliable method to determine the 24-hour energy expenditure (EE) is to
measure the REE via indirect calorimetry. Indirect calorimetry, which measures the
gas exchange [oxygen consumption (V02) and carbon dioxide production (VC02)] at
the IeveI of the lungs, is considered the "gold standardwfor measuring the REE in
hospitalised humans4".
The use of indirect calorimetry for assessing REE in dogs has recently been
examined in veterinary medicine. Prior to 15s. the only clinically feasible method to
assess the 24-hour EE in veterinary patients has been an estimation using the
predictive equation: 70 (BW in kg)'-'=. This value is then corrected for the severity of
illness or injury (PREE k~allday)~"~.The origin of this equation in veterinary medicine
is unclear. Furthermore, there is sparse scientific data to support its use in healthy or
ill dogs. Thus, an alternative method of assessing the 24-hour EE via indirect
caiorirnetry would provide veterinarians with a more reliable method for the
determination of REE.
1.1 GOALS AND HYPOTHESIS
The principle goals of the studies described in this doctoral thesis were as
follows: (1) to confirm the reliability and accuracy of open flow indirect calorimetry in
healthy dogs; (2) to measure the 24-hour REE by open-flow indirect calorimetry in a
group of hospitalised il1 dogs; and (3) to compare the 24-hour REE with the accepted
standard estimates of EE =70 (body weight in kg)'." multiplied by an iilness factor. The measurement of oxygen consumption per kilogram with open-ffow indirect
calorimetry was hypothesised to agree with measurernents of oxygen consumption
per kilogram by closed-circuit spirometry in healthy anaesthetised dogs- Furthemore,
the measurement of REE and oxygen consumption 0(02)via open-flow indirect
calorimetry was hypothesised to demonstrate a clinically acceptable level of reliability
in healthy dogs. Finally, that the estimated 24-hour EE derived from the standard
accepted equation, multiplied by an illness factor, (PREE kcal/day) was hypothesised
to have an unacceptable level of clinical agreement with the measured resting energy
expenditure (MREE kcalrday) for clinical use in il1 hospitaiised dogs.
1.2 LITERATURE REVIEW
1.2.1 Introduction:
Advancements in veterinary medicine have enabled clinicians to successfulIy
treat more challenging and complex medical and surgical problems. Nutritional
support of animals plays an integral role in their recovery. If adequate caloric intake is
not provided during the treatment and recovery phases of illness, the risk of morbidity
and mortality in~reasesl-~~l1-13 . In fact, between 30-50% of hospitalised human
patients are in a state of malnutrition due to anorexia or pseudoanorexia,
malabsorption or an altered metabolic statel 1g13*t4. Moreover, depending on the level
of illness or injury, there wiil be an increase in the REE'~~.Energy and substrates
are mobilised from lean body mass in an attempt to support inflammation, immune
function and tissue repaiP2c27. Adequate nutritional support may help to reduce the
catabolic process and maintain the lean body rnass13n14*1g260281g. However, it may also
be important to avoid excessive intake of nutrients, which is more likely to occur when
aggressive nutritional support techniques are utilised. Excessive nutrient intake may
place additional stress on the respiratory ~ystem~~~~.cardiovascular ~ystem~~and
live$"lnu. Critically il1 humans, with their already compromised respiratory and cardiovascular systems, are more vulnerable to the affects of ~verfeeding~~~~~. Therefore, both under and overfeeding il1 patients can negatively impact on their
clinical o~tcorne~"*~~"*'~~.
It is assumed that hospitalised veterinary patients also have an altered
metabolic state and are therefore at a similar risk as humans to developing protein-
calorie malnutrition. However, sparse data exists to support this ass~mption~*~*~~.
Furthemore, the importance of precisely and reliably assessing the 24-hour EE of il1
or injured animals is pivotal to maximising the beneficial effects of nutritional support.
The purpose of this article is to review starvation metabolism, particularly as
compared to stress hypennetabolism, and to review the sequelea of malnutrition and
overfeeding. In addition, this article will also provide a review of the concept of energy
expenditure with regards to the calculation of energy requirements and the theory and
clinical impact of indirect calorimetry on nutritional support during illness.
1.2.2 Starvation Metabolism:
In carnivores such as the dog, metabolism occurs in a cyclical process. Aftet
a meal, exogenous nutrients are assimilated resulting in a net uptake of glucose
by the liver and free fatty acids by adipose tissue. Assimilation requires an
elevated blood insulin concentration, which remains elevated until the process
has been completed. Once these nutrients have been assimilated, the body
begins to mobilise its endogenous nutrients; the serum insulin levels decrease
and glucagon and growth hormone concentrations increase to provide
substrates for energy metabolism. These endogenous substrates are
predominantly fatty acids and glucose; however, amino acids may also be
utilised for energy depending on the length between meals. If the state of fasting is prolonged, the body will conserve protein at al1 costs and energy
metabolism will eventually be supported by fat oxidation. Eariy in the course of
starvation, the primary fuel source is glucose. The glucose reserves are depleted within 24-hours. Following this, glycerol, lactate, pyruvate and the
amino acids alanine and glutamine are used as precursors for gluconeogenesis8P7738*3g*40.As there are no storage pools of protein present in
the body alanine and glutamine are derived from the breakdown of skeletal
muscle and visceral proteins. Since al1 proteins serve a vital kinction in the
body, excessive catabolism of protein can be life threater~ing'"*'~~'~~~~*~~.
Mortalîty from cardiac or respiratory bilure or from infections secondary to
immunosuppression could occur within 10 daysZ4if catabolism of protein
continues at the rate observed during the early phase of starvation*'.
The ability of many tissues to adapt to an alternative fuel source to decrease
the body's demand for glucose is essential to survival during starvation. Once
the body has adapted to the starvation state, fatty acids and ketones become
the kiel source to al1 but the obligate glucose-using tissues (the brain and red
blood ce1~s)~~-'~*~~~~ . However, the brain will eventually adapt to the ketone
bodies as an energy source. As starvation continues, ketone bodies may accumulate in the blood resulting in metabolic acidosis. In fasting humans the proportion of kiIocalories per day derived from fatty acids is approximately 70 to
85%, from amino acids is 16% and from carbohydrates less than 1o%~'.
Starvation studies performed on dogs9"' demonstrated similar metabolic alterations to those obsewed in humans. In the dog, there is some indication that the glycogen stores may perçist for less than three to four days since these stores Vary with body weight while fasting glucose requirements Vary with metabolic weight (BW . Glucose is gradually replaced by fat as the main energy source in dogs during statvation4'. Fasted dogs will lose befween 1 to 3% of their body weight per day; 56% of this will be due to protein catabolism.
The protein catabolism may be as much as one gram per kg per day in dogs when fasted for 12 days and this rate of protein catabolism lessens as starvation continues'?
Dogs appear to adapt more gradually to starvation than humans. This may be due to the difference in their serum insulin/gIucagon ratio during star~ation~*~.Dogs dernonstrate a mild decrease in serum insulin levels with a mild increase in glucagon levels, which results in normal blood glucose concentrations during starvationa. Hence, fat mobilisation progresses more sfowly in dogs than in humans and the resulting ketotic state is n~ilde?'.~.Dogs do not usually become clinically ketotic during starvation; however, there is an increase in both ketogenesis and ketolysis9". The mild ketotic state maybe a result of the following: (1) there is a restricted increase in the level of free fatty acid concentrations; (2) free fatty acids are directed into a nonketogenic metabolic pathway in the Iiver; and (3) there is minimal suppression of peripheral ketone body utilisation as the serum insulin levels are only slightly decreased. Insulin is required for the peripheral tissues to metabolise ketone bodies in the In fasting dogs. the proporoon of kilocalories derived per day from fat will be approximately 70% to 85%, 15% of which is from ketone bodies. Of the remaining kilocalories 25% are derived from protein catabolism and less than 10% from carbohydrate~'~~~""- This breakdown is similar to the substrate utilisation observed in humans during starvation.
Starvation occurs when there are insufficient nutrients to meet the demands of the body. It is characterised by a decrease in REE, utilisation of an alternative fuel sorirce and reduced protein wasting2023-s26m37p49.These alterations in the body's metabolism are an adaptive process. The goal during starvation is to survive until food becomes available. Therefore, the metabolic response is aimed at decreasing REE and preserving lean body mass7s*37*50. These metabolic adaptations may be completely reversed by providing sufficient exogenous nutrients to meet the body's dernand~'~-~'~*~~~.
1.2.3 Stress Hypennetabolisrn:
In contrast, stress hyperrnetabolism is a generalised response of the body to a variety of stimuli, including shock, trauma, burns, tissue necrosis, infection and sepsis. Energy and nutrients are mobilised to support inflammation, immune function and tissue in an attempt to combat illness or heal injuries1v7*g*'6118m2527740051. After the injury the body enters an initial state of hypometabolism, which persists for 24-48 hours and is referred to as the ebb Stimulation of the sympathetic neivous system, of the hypothalamic-pituitary-adrenal axis and of the infiammatory response during the ebb phase results in the release of catecholamines, corticosteriods, glucagon, growth hormone, cytokines and inflammatory mediator~~*'~~~-~~.The subsequent increase in the REE is referred to as the flow phase. The REE may be increased by either direct or indirect mechanisms. Stimulation of the sympathetic nervous system results in the release of catechofamines, an increase oxygen delivery and consumption at the tissue level and an increase in body core temperature. These three results al1 have a direct effect on increasing the REE'~*=.The REE wiII be increased by 13% in humans for each one-degree Celsius elevation of the core body temperature above normalp. Indirectly, Me REE may be increased by disruption of the Kreb's cycle or oxidative phosphorylation, inefficient metabolism, futile su bstrate cycling of glucose and release of cytokine mediat~rs'~*'~.Catecholamines and counter-regdatory hormones appear to be the primary mediators of the increased
REE. They have the ability to induce glucose intolerance and peripheral insulin resistance through the reduction in the activity of the enzyme pymvate dehydrogenase, which results in the futile cycling of glu~ose'"~~~~~~- Cortisol increases the rate of gluconeogenesis, of protein catabolism in skeletal muscles, and of transportation, by 6 to IOtimes, of free fatty acids and amino acids to the livePg.
Glucagon also dramatically increases the rate of gluconeogenesis and glycogenolysis". Catecholamines (epinephrine and norepinephrine) stimulate both gluconeogenesis and glycogenoly~is'~~~~*~-~~~~~. Furthemore, epinephrine and growth hormone are involved in accelerating lipolysis and suppressing lipogene~is~*~~.~~~~'.
The result of this altered hormonal milieu is an increase in glucose oxidation; however, fewer calories are derived fmm Ïts metabolism due to the futile cycling of glucose frorn lactate or pymvate through the Cori ~ycle'~-'~.As the process of hypemetabolism progresses, the inflarnmatory response may play a more significant role16.18Z.l%5~34 . Cytokines (tumour necrosis factor, intedeukin-1 , interleukin-6) or lipid mediators (platelet activating factor, prostaglandin E2)are released directly from injured tissues or are produced by neutrophils, macrophages and lymphocytes as part of the normal inflammatory response to shock or tissue injury 17*9*161"25*27*37"*51*5555
Cytokine release results in decreased food intake, increased REE, fever and increased acute-phase protein production and is also a potent stimulus for free fatty acid and amino acid mobilisation 7m1601855. Numerous inflammatory mediators have been implicated in heightening the hypermetabolic response. Healthy humans to whom turnour necrosis factor (TNF) and interleukin-1 (IL-1) were administered demonstrated clinical signs sirnilar to those observed after endotoxin administration, such as fever, increased synthesis of hepatic acute-phase protein and activation of the neuro-endocrine pathway? The presence and severity of a hypermetabolic state depends on interactions between the endocrine hormones and inflammatory mediators16.18.40.49
The neuro-endocrine and inflammatory responses result in cardiovascular compensation (increased cardiac output and increased heart rate), sodium and water retention, immunornodulation, initiation of healing and wound repair, hyperrnetabolisrn and an altered substrate metabolism. The body is able to meet the increased energy demands by increasing the oxidation of al1 s~bstrates~~~*'~*'~. Protein catabolism is accelerated beyond the requirements of gluconeogenesis. The mobilised arnino acids are redistributed to support wound healing, cellular inf ammation and hepatic synthesis of acute-phase proteins and to serve as an oxidative fuel source for the penpheral tissue^"-^^*^^^^ . Metabolism of fat for energy increases to such a degree that an essential fatty acid deficiency may become apparent within 10 daysz4?
Unlike the situation in starvation, exogenous glucose, insulin, amino acids or Iipids cannot effectively suppress protein catabolism"*". Malnutrition can develop as rapidly as days in critically il1 humans, resulting in severe protein catabolism of skeletal muscle and visceral organsf6. Currently, no medical therapy has been proven to effectively modulate the body's response to injury or illness and to prevent the redistribution of amino acidssO. Therefore, the aim of nutritional support is the preservation, not the restoration of organ structures and function during the hypermetabolic phase.
In humans, with a single insult and no secondary complications such as, sepsis, wound infection, necrotic tissue or hypotension, the fiow phase will peak in 3-
4 days with resolution in 7-10 days1"18. The responses during the hypermetabolic phase were designed to benefit the organism, enhance recovery and facilitate a retum to health5? If complications occur the neuroendocnne and systemic inflammatory responses are further upregulated and the REE continues to increase potentially resulting in multiple organ failure and death16m1825a51. The intensity of the metabolic derangement and increase in REE depends upon the type and severity of the initial insult and the genetic susceptibility of the h~st'~*~'.Thermal bums and sepsis have produced a hypermetaboiic state in the dog under expenmental
condition^'^*^*. The dogs demonstrated an increase REE, mild to moderate peripheral insulin resistsnce, and an increase in the oxidation of glucose free fatty acids and protein. However, in clinical veterinary medicine the presence or characterisations of hypermetabolism secondary to illness has not been well-d~curnented~~~~~~~. The correction or illness factors used in veterinary medicine, which are meant to account for the hypermetabolic state have been derived mainly from clinical human
Recent investigations, measuring the REE by indirect calonmetry in dogs has demonstrated that there was no increase in REE in a group of dogs in the post- operative or post-trauma period when compared to a group of clinically normal dogsB5, thus bringing into question the assumption that al1 dogs with a similar injury demonstrate the same metabolic response to the injury and a similar increase in the
REE. 9.2.4 Malnutrition:
Malnutrition can result from starvation, inadequate nutrients, problems with digestion
or malabsorption of nutrients or from an increased REE~""~~~.Malnutrition is due
to the depletion of essential nutrients or tissue compartments and if left untreated
results in abnormal body composition and tissue tünction. The alterations of body
composition and tissue fùnction ultimatefy lead to an increased risk of morbidity and
m0mlity1411-13 . Weight loss of 10 to 20% has been associated with functional
abnomalities and poor clinical outcomes"~"*13*14*2626277666667.The prevalence of
malnourished human patients in a general medical and surgical ward has been
reported to be as high as 50%".'~" .This prevalence is likely considerably higher in
critically il1 patients due to the nature of their underlying disease? The prevalence of
malnutrition is, as yet. not well documented in veterinary medi~ine~~.However, the
nutritional support service at the Veterinary Teaching Hospital in Pennsylvania has
reported that 46% of their consultations in a two year period were for the prevention of protein-energy malnutrÏtion and weight loss during ho~pitalisation~~.
The most common nutritional deficiency observed in hospitalised humans is
protein-energy malnutrition, which develops rapidly in acute illness or under stressful conditions. Protein-energy malnutrition has a negative impact on every organ system in the body. In particular, the immune ~ystern'~~~~~~*~~*"n7z, wound hea~ing~~~', the cardiopulmonary system1Z2Bm7641.the gastrointestinal systemsNbZ* , and drug rnetab~lisrn~~are adverçely affectedaB.The most common cause of an acquired immunodeficiency in hospitalised humans is protein-energy maln~trition'*'~.Non- specific, cellular and humoral immunity may al1 be affe~ted"~~.Non-specific immunity is impaired due to a marked decrease in the bactericidal capabilities of neutrophils and macrophages and a decrease in the serum concentration and function of complement proteins2.Cellular immunity may also be flawed due to a decrease in the number and activity of lymphocytes and significant atrophy of the thymus, spleen and peripheral lymph no de^'"^". Finally, there appean to be a lack of reactivity to cornmon recall antigens (anergy13 and a decrease in the binding affinity of antibodies
to antigen~''~~.
Malnutrition may result in decreased activity of the earfy wound fibroblasts,
thus retarding the normal healing processT4and resulting in a decrease in the early
wound strength. The ultimate strength of the wound, however, is ~naffected'~*'~.
The intestinal tract walts and the Kuppfer cells in the liver protect the host from
intestinal intraluminal bacteriaa3.If there is an alteration in the thickness or
perrneability of the intestinal epithelium, as occurs with protein-caloric malnutrition,
haemorrhagic shock, cancer chemotherapy or direct trauma, then bacteria are able to
translocate into the portal blood stream6-47"283.The Kuppfer cells in the liver act as a
back-up bamer to prevent the systemic invasion of these microbes. However, when
these cells become ovenivhelmed, they release interleukin-1, which results in the
upregulaüon of the hypercatabolic response". The sequelea of protein-calorie
malnutrition include an increased risk of septic complications secondary to the
systemic invasion of intestinal bacteria, an upregulation of the hypercatabolic
response, poor wound heaiing and an acquired immunodeficient state.
It is a misconception that malnounshed patients are only prone to an excess
risk of septic complication2. In fact, they are abo at significant risk of cardiopulmonary
complications due to the decreased strength of the heart, diaphragm and inspiratory
and expiratory muscles that occurs secondary to the catabolism of the muscles'
amino acids and depleted fuel supply2. The resulting marked cardiac muscle atrophy,
decrease in left ventricular contractility and altered cornpliance causes a decrease in
cardiac The respiratory muscles demonstrate a decrease in their
endurance and may be unable to respond to the increased ventilatory demand
associated with a hypercatabolic state3. In fact, moderate semistarvation of
approximately 7 days will decrease the ventilatory response to hypoxic stimuli and to
a lesser extent to hypercapneic stimuli79. These alterations increase the risk of a
' patient developing respiratory failure. Finally, the normal respiratory defence mechanisrns are irnpaired. There is a decrease in the antioxidant defence
mechanism, alterations in surfactant production, an ineffective cough response and a
suppression of pen'odic deep breaths. These impaiments al1 contribute to an
increased susceptibility to respiratory infections.
Drug metabolism and pharmacokinetics are also affected by the nutritional status of the patienf? The oxidative reactions are more susceptible to nutritional effects than are the conjugation reactions. Therefore, protein-calorie malnutrition may decrease the clearance of many oxidative drugs by about 20 to 40%'~.
1.2.5 Overfeeding:
In human medicine, earlier and more aggressive nutritional support was advocated to compensate for the significant nutritional deficiencies present in disease. In the 1970's and 1980's, many human patients received excessive calories and nutrients The administration of excessive carbohydrates was associated with; hyperglycaemia; osmotic diuresis; increased REE~~*~*~~*~*~~~'*~~~~~;increased concentrations of orga nic acidss; and respiratory and hepatic dysfunction30"1"~38~8s~go.
Excessive lipid intake was associated with: hypedipidemia3'*52;alterations in immune
an increased rate of infection'^^^; promotion of oxidant-induced ce11 injuryJp; and compromise in pulmonary diffusion capacity"*". Human medicine has now recognised that there are few advantages to overfeeding"; moreover, overfeedingw6 is frequently associated with an increased n'sk of morbidity and rn~rtality~*~'~*~.Thus, a concept of moderation has developed in human medicine that places an increased emphasis on matching caloric intake to energy expendit~re~~~~*"~'"8.
1.2.6 Daily energy expenditure:
Energy is neither created nor destroyed. Therefore, an animal must continuously transform energy by oxidation of both endogenous and exogenous fuels to produce energy, heat and mechanical work. The goal of ingestion of food (caloric intake) in an individual is to meet their daily EE? Therefore, in il1 individuals who are unable to correctly balance their food intake to meet their 24-hour EE, an estimate or
measure of the 24-hour EE is requiredsT.The use of predictive equations with
correction factors for physical activity and iilness have proven an unreliable method of
estimating 24-hour EE in critically il1 humans35gQ7~QQ. A more reliable estimate of the
24-hour EE is to derived the measurement of gas exchange with an indirect
caloimeterat the bedside21.~.53.84.87.98s100-102 . However, the standard of tare in
vetefinary medicine continues to be the use of predictive equations, corrected with an
illness factor, to estimate the 24-hour EE~-~~*~~-~~~~~~~~*'~~'~~.
The energy required for the maintenance of ion pumps, for synthesis and
degradation of cellular constituents, for biochernical cycles, for the electron potential
gradients across membranes and for the function of the liver, kidney, heart and brain
constitutes approximately 70% to 80% of the 24-hour EE~~~~~~~~~*~~'~~~~*~~~. The overall
metabolic rate is deterrnined by the energetics of the chemical reactions in these
tissues and by the interactions between the sympathetic nervous system and thyroid
Historically, these energy-requiring processes have been referred to as
basal energy expenditure (BEE). The BEE was measured in hurnans under very
restricted conditions: after an overnight sleep; first thing upon awakening in the
morning or when still asleep; after at least a 1O hour fast; in supine position with a
normal body temperature in a thermal neutral environment; and without physiological or psychological These conditions are impractical to obtain in a hospitalised setting in either humans or dogs. Therefore, a more clinically applicable measurement of EE has been suggested, the resting energy expenditure (~~~)23.'OQ.i13,ii4. Resting energy expenditure may be defined as the EE measured in an individual at rest in a thermoneutral en~ironment'~*'~~.The major difference between the REE and BEE is that the former is inf uenced by the increase in metabolic rate secondary to ingestion of food"s and the conditions under which it is measured are less rigomusly defined? Moteover, REE may include components due to physical activity, psychological stress or variations in the body or ambient temperaturea7. In animals, it is irnpractical to completely control al1 of the spontaneous
muscular activity or psychological stressors that may be present during the testing. A
true post-absorptive state rnay also be impractical to achieve114. Therefore, REE is a
more feasible measurement of the minimal metabolic rate in animals.
The metabolic process of balancing the body's energy demands to supply may
be influenced by a large number of factors: gender, age, environment, nutritional
status, tirne since last meal, physical activity, time of day and psychological
conditions. The sum of these factors plus the energy demands of REE rnake up the
24-hour EE. Several of the more general factors that influence 24-hour EE should be
ernphasised, including gender, age, environment, body ternperature, physical activity,
and diet-induced thermogenesis. Energy expenditure is proportional to the number of
metabolically active cells in the body. The vast majority of these cells are located
within lean body tissue. Therefore, as women have a lower percentage of lean body
tissue than men, a gender difference in REE is present in h~mans'~~'~'-~~~.This
gender difference in REE does not appear to be evident in dogs. In reports of hnro
feeding trial studies, with 78 and 94 adult dogs respectively, no significant difference
in EE was observed based on gender116*117.It should be noted that the gender
comparison in these studies was between intact male and female dogs.
The REE in humans and animals varies with age8'. lncreasing age is
associated with a decrease in REE due to a decrease in the Iean body mass and an
increase in the total body fat store^^^^"^. In dogs, EE was infiuenced by age (ranged frorn 8 rnonths to 12 years) in one study"'. However, in a second study, which examined fasting EE in 94 adult dogs between the ages of 1 to 6 years, no significant difference due to age was observed"'. It seems probable that in the second study the age range was insufficient to demonstrate a difference in EE.
The temperature of the environment exerts a large influence on the 24-hour
EE. However, if an individual remains in their thermoneutral zone, there will be a negligible effect on the overail REE. The thermoneutral zone is defined as the ambient temperature above or below which the REE of a nude subject begins to rise in order to maintain body temperature and heat balance within normal limits8'. The therrnoneutral zone in the dog is reported to be 18 to 25 degrees els si us"^. When the ambient temperature increases, there is a concurrent n'se in the REE due to increased energy losses via sweating or pantingB7.Panting dogs have an increase in
REE due in part to the increased oxygen consumption by the respiratoty muscles120.
Cool environments or hypothermia also increases REE through shivering and non- shivering therrnogenesi~~~~~~~~~~. In dogs, REE can increase by 30% after 20 minutes of shivering. If the ambient temperature continues to decrease to 12OC, the body core temperature cxrnnot be maintained and a significant decrease in the REE will be obse~ed.Anesthezied, paralysed and ventilated dogs that were cooled to a rectal temperature of 29 OC, demonstrated a decrease in their REE of 52%B7.ln hospitalised patients, the environmental temperature tends to be controlled such that they rernain within their thermoneutral zone resulting in Iittle overall affect of the environment on
REE. However, humans with rnoderate to severe bum injuries demonstrate a higher zone of thermal neutrality; if their ambient temperature is not increased to 32°C they experience a significant increase in 24-hour EE~~*~~~It is unknown, as yet, if dogs alter their zone of thermal neutrality in response to disease.
Body temperature is regulated by complex mechanisms, which starts at the level of the peripheral receptors and terminates at the level of vasoactive reactions.
These complex mechanisms result in the modulation of blood flow and therefore heat loss to the skin? Body core temperature and thermogenesis have a significant effect on REE such that for each one degree Celsius increase in core temperature there is a
13% increase in REE"~"~.The opposite effect is obsewed when the body core temperature decreases.
The ingestion of food has long been noted to have an effect on REE"~-'~"*'~~.
The increase in heat production following a meal occurs in two phases. The first phase, which is due to the stimulus associated with the palatability of a meal, is referred to as the cephalic phase. The second phase is associated with the actuat cost of digesting, absorbing and storing the nutrients. Dogs are very responsive to palatable food; thus the cephalic phase may last up to 45 minutes in this speciesl".
LeBlanc et al1= demonstrated that dogs have an increase in V02in the post-prandial penod. The degree and duration of this increase in V02 depends on the size of the meal and more importantly the on frequency of the feedings. A large meal of 500 g had a sustained effect on the V02and it took 6 hours for the REE to retum to initial values. A small meal of 125 g demonstrated a less sustained increase in V02and the
REE returned to the initial value within 3 hours. However, dogs fed a large meal as compared to the smaller meal demonstrated only a 3% difference in the 24-hour EE.
This difference would be difficult to detect with indirect calorimetry"? Furthemore, the post-prandial REE was twice as large when the meal frequency was increased'".
The increased REE was attributed to repeated sensory stimulus associated with the palatable foods. These studies illustrate that there does appear to be an effect of diet- induced thermogenesis on the REE in the dog. It appears that the frequency of feeding has the largest effect rather than the meal size on the 24-hour EE'? Diet- induced thermogenesis depends not only on the size and frequency of a meal, but also on the following factors: the composition of ingested food, the time elapsed since the last mea!, the nutritional status, the underlying disease process and the quantity of nutrients provideds7. If nutrients are provided below actual EE then there is no discernible increase in the REE. However, if nutrients are provided in excess, particularly in the form of carbohydrates, then the REE maybe increased by 25 to 300h34.5W7.123-125
The addition of an activity factor to the REE in the calculation of 24-hour EE is recommended to account for physical activity. In bed-ridden humans, and presumably in cage-rested dogs, the 24-hour EE is almost identical to the extrapolated 24-hour
REE~~.However, significant increases in the REE may be associated with routine nursing procedures such as bed baths, dressing changes or repositioning. Even the spontaneous movernent of a body limb may result in an increase in the REE by
approximately 10 to 15%124*126 . These activities result in a transient increase of the
REE and do not contribute significantly to the 24-hour EE"~*'".
Depth and duration of sleep affects REE and is most Iikely responsible for the
observed diurnal variation in REE? In normal humans, a cyclical difference in
oxygen consumption is observed between day and night, with the Vonbeing
approximately 20% greater during the day than at night127*128-Van Lanschot et ali2'
demonstrated, in a group of critically il1 mechanically ventilated patients, that there
was very little difference in the REE throughout the 24-hour period. This particular
group of patients did not demonstrate a typical diurnal variation in the REE. This may
have be due the greater quantity of interactions with the nursing staff, therapeutic
interventions and diagnostic procedures, which left Iitüe time for rest or sleep. As their
condition improved, less care was required and the normal diurnal variation again
became evident118-lllness and hospitalisation rnay have a significant impact on the
normal diurnal variations observed in the REE. The measurement of REE by indirect
caloflrnetry is at its initial stages in veterinary medicine. Currently, veterinarians do not
have any data on the influences of hospitalisation or illnesses on the normal circadian
rhythms of REE in dogs65~130-132. However, significant variations in the VOz have been
associated with the level of alertness in the dog. Peters et perfomed open-
flow indirect calorimetry rneasurements on trained dogs for several hours under
standard conditions. The observed VOzvalues fluctuated from 2 to 1O ml/min/kg in the
resting state. The mean V020f 5.57 2 1.18 (SD) mllmidkg in the alert state
decreased to 3.97 + 1-01 ml/min/kg after the dogs became drowsy. The V02 at its
lowest point, with the least variation during natural sleep (2.46 5 0.48
ml/midkg).Therefore, perforrning multiple measures of the REE in a 24-hour period with the subject in a similar level of alertness is important as it provides a more reliable estimate of the 24-hour EE~~-~~."~~*~~~. lllness and injury tend to increase REE"'; however. not al1 critically il1 humans demonstrate the classical hypennetabolic response to illness. Several
StUdieSg7.10î.':18,137,138 have demonstrated that in 35 to 65% of patients a hypermetabolic response was observed, whereas 30 to 50% were normometabolic and 15 to 20% were hypometabolic- Therefore, a particular diagnosis shoutd not automatically be associated with a hypemetabolic response in humans. Curent research in veterinary medicine has provided evidence to support this statement in a group of postoperative and traumatised dogs. The mean RE€ in the group of injured dogs was not significantly different Rom the mean REE in the group of healthy dogs.
The REE was measured first thing in the morning (05:OO to 08:OO) at 18 to 72 hours after the injury or surgical procedure by open fiow-indirect calorimetrys5.
Medications may significantly influence the REE. Catecholamines, pressor agents or thyroxine supplementation increase the REE, whereas sedatives, opioid analgesics, beta-blockers, alpha-blockers and general anaesthesia al1 reduce the REE inhumansl18.124.126.139 . General anaesthesia rnay decrease the RE€ in humans by 10 to 20%"'. However. the effects of general anaesthesia on the REE in dogs are not as cleariy defined. The effects of surgical trauma and general anaesthesia on the
REE have been assessed with indirect calon'metry in 40 dogs undergoing a variety of surgical pro ce dure^'^'. The preoperative EE was compared to the postoperative EE at 24,48 and 72 hours and at 14 days after the surgical procedure. The surgical procedure did not significantly alter EE and there were no obvious effects of anaesthesia. Elliott et al1% noted that anesthetized V02values in dogs were decreased relative to the resting values in an alert state but were increased relative to those observed during natural sleep. Therefore, in the dog, general anaesthesia may not autornatically indicate an absolute decrease in the REE and the effects may depend upon the anaesthetic agents uti~ised'~~*'?Sedative and analgesic medication decrease the REE in humans by 12.7 to 15%1188124 . The effects of these medications on EE have not been evaluated in veterinary medicine; however, we assume that the physiological reactions of dogs to these medications are similar to those observed in
humans.
Energy expenditure may also be inftuenced by differences in individual
metabolic responses to injury. A wide and variable range of responses to illness or
injury are observed in critically il1 hum an^^^^^.". These differences are apparent
despite similarities in signalment, severity of illness or injury and therapy. They are
likely due to inherited traits such as genetically determined differences in skeletal
muscle metabo~ism'~~.
Artifacts may be introduced into the rneasurement of REE, particulariy when
performed on patients that are not in a resting or a metabolic steady state140. Gas
exchange measurements quantitate the carbon dioxide elimination and oxygen
uptake; we assume that these are equal to the VC02 and Von.However, this
assumption is npt valid if the body is not in a state of rest or is able to quickly alter its
store of oxygen and carbon dioxide gases. The body's gas stores are affected by
alterations in the minute ventilation, which may occur secondary to pain, anxiety,
subject discornfort, hyperventilation or hypoventi~ation'~~.Pain, agitation and
restlessness may contribute up to 10% of the 24-hour EE in critically il[ humans
IZ4ill8187. Therefore, rnetabolic measurements should not be performed within 30 to 60
minutes of a painful, therapeutic or diagnostic pro~edure"~~'~~.Errors in measurement
of these gases, wilI occur if there are any Ieaks in the system; volume losses of both
carbon dioxide and oxygen will artificially decrease the REE"~.'~'.
1.Z.7 Estimated energy expenditure:
As early as the eighteen-century Lavoisier noted that large animals produce
more heat than smaller ones. Subsequently it was demonstrated that smaller animals
produced more heat per unit weight than larger animal^"^. The measurement of metabolic rate requires a base, which takes into account the differences in body size within and between ~pecies'~~*"~.In 1883, Rubner performed the first metabolic experiments on dogs using direct and indirect respiratory calorimetry under standard conditions of fasting, resting and in a thermoneutral en~ironment'~~.These
experiments demonstrated that heat production per unit of body surface area was
independent of size. These findings were then confimed in dogs by ~usk'~,Kunde
and c te in ha us'^^, DeBeer and Hjort lUand in a variety of other species by
This relationship of body surface area to metabolic rate became known as the
Surface ~aw'~',which states that large and small bodies of similar shape have a
surface area in proportion to the square of their linear dimensions or a 2/3 pawer of
their body weightlM*"'. This mathematical relationship assumes that within a species
the animals are homoIogous in body composition and shape; however, this
relationship does not appear to hold true for al1 species. ~oczopko"~in 1971
demonstrated that if body surface area, is rneasured in mice, pigs, and horses and then compared to the formula used to predict body surface area the results differ by
130% in mice, 70% in pigs and 30% in horses. Kleiber also noted that surface area was an illdefined concept and poorly standardised in animals. Data From his laboratory indicated that the metabolic rate per unit surface body area had a tendency to be greater in larger animals than in smaller ones. This indicated that the metabolic rate expressed in terms of body surface area was not independent of size in1 al1 animals as previously assumedl?
Intuitively, it makes sense that the rate of heat toss frorn a body to the environment would be proportional to the body surface area. However, there is no theoretical basis for the hypothesis that the metabolic rate of homeotherms shoutd be exactly proportional to the body surface area, rather than to a more general uneasure of body sizelo4.Kleiber in January 1932'~and Brody and Proctor in April 19321°5 measured the fasting metabolic rate of mature animals in 26 species in a respiratory calorimeter under standard conditions. They noted that there was a mathematical relationship between the logarithm of body weight to the loganthm of fasting EE. The experimental data indicated that the metabolic rate was proportional to a power of the body weight greater than the two-thirds power that had been previousiy defined under the Surface ~aw~~~~~~~. Kleiber found that the mean standard metabolic rate of
mammals was 70 x (BW in whereas Brody's equation was 70.5 x (BW in
104*105. ln 1964, the European Association for Animal Production selected by
convention the allometric equation of the body weight in to compare
intraspecies and interspecies metabolic rates". Kleiber's equation was selected since
it was simpler to use and could be calculated without the use of logarithms.
Moreover, Kleiber's and Brody's equations agreed when compared over a wide range
of weights frorn a 1 0-gram mouse to a wha~e~~~"~."~. In vetennary medicine, the
Kleiber predicative equation is considered the standard and is presently utilised for
estirnating the REE~-~~~*"~~~~'~~.Given. the 100-fold variation in body size observed
across the vanous breeds of adult dogs, the Kleiber equation may be thought to most
precisely estimate the REE in dogs'03. However, the reason that the Kleiber equation
was generally adopted in veterinary medicine, instead of using specific estimates
based on experimental data of REE in the dog, rernains uncleaPsa. Dogs
demonstrate a wide range of weights when mature; however, this range is small when
compared to the range used in the denvation of Kleiber's equation. Kleiber's equation
is the most reliable when comparing metabolic rates over large ranges of body weight. When these ranges becorne smaller, the exponent may range from 0.6 to
o.gl03.142.144-146 . ~eusner'~~,reviewed metaboiic data from the medical Iiterature in 332 mature healthy dogs, where the EE was measured in a fasted, rested state in a thermoneutral environment using calorimetry. This data represents the best available approximations of REE in the dog to date. The data revealed that in the dog a single regression line did not accurately describe the relationships between the logarithms of fasted EE and body weight. The exponent of 0.75 was not accurate in the dog data and the average regression line was more accurately fitted by a metabolic power function of 0.86~~"~~.However, there was still significant individual variability in the relationship between the body weight to this 0.86 power function and the fasted EE.
This variability was suficiently great that the average regression line (70x BW O-") could not be used as a reasonable estimate of the REE in an individual dog. By focusing on the rnass exponent veterinarians will either be overfeeding or underfeeding a significant percentage of healthy dogs, let alone il1 or hospitalised d~~s~*'*~.It seerns unlikely that the current predictive equations can yield a clinical acceptable estirnate of the REE in hospitalised dogs.
In healthy people, the REE as predicted by the Harris-Benedict equations was within 10% of the measured REE by indirect ca~orimetry~~~'~~.Thus, demonstrating that there appears to be little variation in the resting metabolic rate in nomal humans"'. However, these equations have recently been re-evaluated in healthy humans with modem indirect and direct calon'rnetry techniques. In these studies, the predictiie equation, on average, overestimates the REE by 12.3%. Moreover, the predictive equation could underestimate the REE by as much as 14.7% or overestimate it by as much as 19.1 %148.Therefore, the Harris-Benedict equation is not as accurate in the estimation of the REE in healthy humans as was previously thought.
Traditionally, the Hams-Benedict equations, with correction factors applied, which account for the degree of hyperrnetabolism and physical activity, have been utilised in clinical medicine to estimate the REE of il1 and healthy indiVidualS6'",9'.124.137,149.1~ . Long et alm detennined REE by indirect calorimetry for a group of patients following elective surgical procedures, skeletal trauma, skeletal trauma with head injury, blunt trauma, sepsis and burns. The predicted energy requirements were calculated using the Hams-Benedict equations and then adjusted upward with activity and illness factors so that they agreed with the average measured REE in each group. These adjustment factors are the bases from which the standard iliness or stress factors have been derived. Long et al demonstrated that there was an increase in the average REE in al1 groups; the increase in REE correlated with the degree of injury. However, the variation in the measured REE of the individual patients was so great that the rnean estimates of the groups could not be used to predict, with any statistical assurance, the actual REE for a particular
patient. Furthemore, many investigators have demonstrated that the use of predictive
equations, in the i~dividualcritically il1 human patient, is inappropriate for an accurate
estimation of the ~~~21.35,Q7*9.1".137.148.'50-152. TO predict the REE the height and
weight of an individual must reflect the metabolically active portion of the body?
Therefore, part of the inaccuracy of these equations may be due to alterations in the
actual weight and body composition subsequent to pre-hospitalisation illness,
extracellular volume alterations and fluid resu~citation'~*'~'. The predictive equations
assume that al1 patients respond to iilness and injury in a standard manner. This
concept discounts the presence and significance of biological variability. There are
likely significant differences in the inherited patterns of metabolic responses amongst
patients resulting in differences in REE despite similar injuty6. A more reliable method
for the determination of REE would be to measure the REE by indirect calorimetry
severaltjmes a day21,2TT35,97-98,1O2,l24.l 37.148.1 50-1 52
Currently, in il1 dogs, the predictive REE is expressed as 70 (BW in kg)
and is multiplied by an illness factor. The illness factors are subjective correction
factors, ranging from 1.25 to 2.0, which are selected by the clinician and are based on
the severity of the injury or il~ness~-'~.The illness factors, in veterinary medicine, have
been denved from average metabolic measurements performed in populations of
human patients with sirnilar disease processes and are intended to correct the REE for hypermetabolic responses to injury"". However, not al1 patients demonstrate a classic hypemetabolic response to injury. Furthermore, the magnitude of the REE increase is not as great as previously expected in humans98~'02'24w'58.To date, there is liffle scientific data to support or verify the effectiveness of the illness factors in il1 or injured dogssJ. Therefore, further studies measuring the REE by indirect calorimetry in hospitalised dogs, with a vanety of illnesses, are required to assess the effectiveness of the predictive equations. Walton et al 65 have recently provided some scientific data to clarify this controversy. They demonstrated that the mean RE€ in postoperative or traumatised dogs was not significantly greater than the mean REE in healthy dogs,
suggesting that an illness factor may not be required.
1.2.8 Measured Energy Expenditure:
Fire, as a metaphor for life, has a basis in biology. Lavoisier, in the mid-
eighteenth century, was the first to point out that the combustion of a candle and the
respiration of an animal involved the removal of a gas from air. This gas was termed
"oxygene" and it was replaced in air with a gas containing carbon. He further
observed that there was a fixed relationship between heat production and the gas
exchange of oxygen (VOz) and carbon dioxide (VCOp),both in metabolism and in
fire154.155 . During metabolism, humans combust fuels akin to a fumace buming coal.
These fuels, carbohydrates, protein and lipids provide chemical energy. The potential
chemical energy in these molecules is liberated in the body through oxidative
processes, which utilise oxygen and produce the end products of carbon dioxide and water. During the process, some of the transferred energy is lost as heat and the remainder is trapped as ATP, which may then be transported to areas of the body where energy is req~ired'~~.'".Eventually, however. al1 energy will be converted to heat during a resting statel-.
Calorimetry is a method that measures energy by measuring the production of heat'58v'59.Calorimetry is based on the Laws of Thermodynamics: (1) energy can neither be created nor destroyed but can only be exchanged between the body and the environment; and (2) any change in the total energy content of a system (e.g., the heat of combustion in biologic oxidation) results in a change in both the free energy and the entropy (randomness) of the system 53~153*'"*161 . The second principle of thermodynamics is illustrated by the following equation: AH= AG + TAS, where AH known as enthalpy, is equal to the total heat content of a substance or physical systern; AG know as free energy, is equal to the orderly energy, capable of performing work; and T is equal to temperature; AS known as entropy, is equal to the degree in which the total energy of a system is uniformly distributed (randomness) and thus unavailable to do work. Entropy is maximised in spontaneous running
systems such as biochemical oxidative reactions. As the oxidation of carbohydrates,
lipid and protein occurs, there is an increase in the entropy or randomness of the
system, which results in a decrease in the amount of energy transferred to ATP.
Therefore. energy-yielding reactions are invariably less than 100% efficient 1n.161. The
measurement of metabolic free energy conversion or energy expenditure is
perfonned using direct or indirect calorirnetry. Direct and indirect calorimetry
examines each side of the heat balance equation: the rate of heat removal from the
body and the rate of heat production of the body respective~y'~.
Direct calorimetry is considered the gold standard for measuring EE over long
periods of tirne and is able to measure the total heat produced by the body accurately and pre~isely~*'"*'~~.The principle is simple; however, its application in a clinical situation is difficult, as it requires placement of the subject into a specially constructed r~orn"*'~.Indirect calorimetry is accomplished by rneasurement of the products of oxidative metabolism. Respiratory gas exchange measurements such as oxygen consumption 0/02). carbon dioxide production 0/C02), minute ventilation &), and the respiratory quotient (RQ) are detemined from expired gas analysis. The gas exchange measurements. VOz and VC02 ,are used to estimate the REE, expressed as kilocalories per day by the method of Weir or ~usk~~~'~~~'~~.The RQ (VC02/V02) may also be used to detennine the type and rate of substrate oxidation in
Viv053.1 l8.lZ6.lS7.158 . cellent agreement, with differences of less than 1%, between the two calorimetry methods are observed both in humans and dogs when in a resting
Steady state 108.1 13.16"16* 25
1.2.9 Theoretical Considerations of Indirect Calon'metry:
ATP FUEL+ OXYGEN C@+Hi O+HEAT
RESPIRATORYQUOTl ENT
Figure 1.Y. Theoretical considerations in the application of the indirect caloflrnetry technique for the calculation of resting energy expenditure.
Indirect calorimetry estimates body heat production or EE by measuring the whole body V02, VC02 and VE. The gas exchange measurements, VOpand VC02. performed by indirect calorimetry must be converted from the units of litres into kilocalories. The energy equivalent of V02varies with the RQ (VC021 V04; therefore, the amount of heat released per litre of oxygen consumed depends upon the nutrient that is o~idised'~~.The conversion from litre to kilocalories may be accomplished by calculating the non-protein RQ from the gas exchange measurements. The non- protein RQ is then used to assign a caloric value to the consumed oxygen based on standardised tab~es~~*'~~-'~~.The second method is the application of the Weir equationln*'", which is the accepted standard equation used to calculate the REE.
The REE = [(3.94)0/02) + (1 .l1)(VC02)] 1.44 - (2.17 UN) where RE€= kcalfday, V02
= oxygen consumption in ml/min, VC02= carbon dioxide production in mllmin, UN= total urinary nitrogen in grams Iday (urea + non-urea nitrogen) and 1.44 = 1440
(min/day)/l000 (mlllitre). The Weir equation has demonstrated close agreement to direct calorimetric assessrnent of EE and is able to depict the relative importance of each parameter in the estimation of REE'~~.A 5 % error in VOz for a given RQ, will produce a 3.5% error in the REE; however, the same magnitude of error in the VCO2 will decrease the obseived error by three to four timesla. The impact of an error in total urine nitrogen is of Iess significance, as a 100% error will only produce a 1% alteration in REE'~~*'~?T~~abbreviated Weir formula, which excludes the urine nitrogen is more frequently applied to calculate the REE, as a 24-hour urine collection may not always be feasible. The abbreviated Weir formula = [(3.94)0(02)+
(1.1)(VC02)] x 1.44 and is expressed in kilocalories per day 159.180 .
Indirect calorimetry continuously measures the oxygen and carbon dioxide concentration in respiratory gases as well as the fiow rate of these gases. The amount of oxygen removed from inspired air is not always equivalent to the amount of carbon dioxide exhaled due to the tissue level differences in oxygen consumption and carbon dioxide production. Therefore the inspiratory and expiratory gas flow rates and volumes are not identical. Technically, it is very difficult to measure inspiratory gas volumes without significant error. The measurement of inspiratory gas flow rates can be negated by the application of the Haldane transfornati~n~~*~~*~~~*~~~.There are several assumptions inherent in the use of the Haldane transformation. The transformation assumes that the wmposition of room air is faitly constant; oxygen:
(O2)=20.93%, carbon dioxide (COz)= 0.03% and nitrogen (N2)=79.04%. Since the percentage of carbon dioxide in room air is negligible, it is considered to be zero.
Furthemore, the inspired and expired gases are only composed of these three gases and nitrogen is considered an inert gas in terms of metabolism, as the body is unable to metabolise or produce this gas. This indicates that any difference in the concentration of nitrogen between inspired and expired air is due to a difference in the number of O2 molecules consumed by the body, which were not replaced by an equal number of CO2 molecules produced during metabolisrn. It is this difference between the inspired and expired nitrogen concentration that allows the volume of inspired air to be determined hmthe volume of expired air using the Haldane
transformation.
Indirect Calorirnetrv Euuations:
1) Vo2=[Vl~F1o2]-[VExFEO2]
2) VCO2= [ VE x FECO2 ] - [ VI x FIC02] ( Note: FIC02 is considered to be zero)
3) VI = VE x [FENZ/ FIN2 ] : Haldane Transformafion
4) FEN2= 1-[FEOZ-FEC02]
5) FI NZ = 1- FiOz
6) VE(STPD) = VE(ATPSI x [PB- P~20/ 7601 x [273 / 273 +
7) VO2= [( 1- (F~02- FEC02 ) X Fi02) - FEOZ] X VE(STPD) X 1000
ATPS, ambient temperature, pressure saturated; PB, barometric pressure (mmHg); CO2 carbon dioxide; E, expired air (rnl/min); F, fraction; I inspired air (ml/min): Na, nitrogen; 02 oxygen; PH2O, partial pressure of water vapour in a saturated atmosphere at T; STDP. Standard temperature and pressure, dry; T, temperature ("C); V, volume/time (min); VCOZ , carbon dioxide production (ml/min); V02, oxygen consumption ( ml/min).
Haldane developed these equations in the 1920's and most commercial instruments still utilise this transformation when assessing VOZ53*'65-1e9g'70. There are limitations in using the Haldane transformation. In vitro studies have demonstrated that the error in
V02 measurements increases significantly with inspired oxygen content, particularly as it exceeds 0.6-0.8~~*'~~*'~~~~~'.AS the percentage of oxygen inspired increases and approaches one, the difference between the inhaled and exhaled nitrogen content decreases and the denominator in the above equation approaches zero. Therefore. any slight error in measuring the expired or inspired oxygen concentration would result in large errors in the determination of oxygen consumption and therefore EE. In order to Iimit the error in the determination of V02the percentage of inspired oxygen must be measured to the second decimaf point or al1 supplemental oxygen discontinued during the measurement peri~d'~~."'.
There are a variety of techniques for the measurement of expired air volume and gas concentrations (FEO2,FECOZ) utilised in indirect calorimetry. These include:
Douglas bag collection system, Tissot spirometer, fiow-through mixing chamber and breath-by-breath systems"'. The ~erosport~TEEM 100 metabolic analysis system utilises a modified breath-by-breath system in order to analyse the expired gas concentrations. The calorimeter analyses the expired gas concentrations in each breath; however, these results are averaged over a 20 second periodl".
1.2.10 Clinical Application of Indirect Calorirnetry:
To correctly interpret and apply the results of indirect calorimetry to an individual patient, a clear understanding of the assumptions and principles on which the technique is based is required12g157.The first assumption is that al1 oxygen is completely and rapidly utilised in oxidative metabolism. The second is that al1 expired carbon dioxide is derived only from the cornplete oxidation of fuels without any transit delay between the cells and expired gas. The third assumption is that al1 the nitrogen resulting from the oxidation of protein is precisely measured in the urine. The latter assumption may be omitted in the clinical setting since a less than 2% error in the 24- hour EE would occur if the nitrogen excretion was excluded and the estimate was based solely on gas exchange measu~ements'".'~. The first assumption is fairly simple to satisfy as O2diffuses rapidly at the cellular Ievel and there is a small pool of
O2inthe body (1.2 L) compared with its global flux of 300-400 Uday in humans. This means that any alteration in the oxidative phosphorylation process results in an equivalent alteration in the ~02'~.The second assurnption is not as easily satisfied.
This is due to that fact that there is a relative large body pool of COz (1 5 L)'~~ compared to its global flux of 300 Uday in humans, Therefore, a delay would be expected pnor to a change in the expired CO2. The second assumption may only fully be met if the patient is in a metabolic steady state. The steady state ensures that gas exchange measurements are equal to tissue gas exchange, for example the oxygen uptake will be equal to VOz and carbon dioxide production will be equivalent to
VCO:". This state is present at rest or if there are no acute alterations in the ventilatory pattern, the acid-base status or the metabolic production of carbon dioxide.
An important consideration in obtaining a steady state REE reading is the duration and frequency of the metabolic readings in a 24-hour period. Van Lanschot et alz1 demonstrated, in mechanically ventilated critically il1 hurnans, that two 15-minute readings 12 hours apart were sufficient to give an estimate of the 24-hour REE to within 4% of the actual24-hour EE. Recently, several investigators have suggested that the extrapolation of five one-minute measures of the REE would be suficient if the coefficient of variation of the minute-to-minute measurement was less than
~~~136.158.1sS.172.173. In most instances, the mean value of the REE during the brief reading was not significantly different from the rnean value of a longer pei-iod of rneas~rernent'~.The larger the variation in the minute-to-minute readings, the longer the measurement time required to obtain a reliable REE measurernent. For example, if there were a 10% coefficient of variation in the minute-to-minute readings, then a
1O-minute steady state reading would be required to assure a reliable value of the 24- hour EE". A steady state measurement should be obtainable in a 15 to 30 minute reading and should be identified according to statistically defined g~idelines~~~~~-"~.
Unfortunately, little work has been done to evaluate these issues with the clinical use of indirect calorimetry in the dog. A recent study by Walten et alq3'has dernonstrated that indirect calorimetry is a reliable measure of REE and V02in healthy client-owned dogs. An interclass correlation coefficient of 0.87 was achieved when the rniddle three readings during the day were averaged. The investigators
noted that the first REE reading was significantiy greater than the subsequent
readings and should be discounted. Furthermore, the reliability of the REE was
increased with more frequent readings and not by increasing the duration of the
rnea~urement'~~.Therefore, large errors in the 24-hour EE of a dog could be obtained if only one rneasure of REE was perf~rrned'~~This has also been dernonstrated in healthy hum an^'^^. Moreover, there are several technical aspects that must be taken into account to ensure reliable measurernents. The metabolic apparatus must be portable and contain the foilowing: accurate and sensitive O2and COnanalysers; a device to rneasure the expired volume or flow of air, temperature, barometric pressure, time; and a rnethod to dry the expired air. Nso, the machine must contain the appropriate software to store data and perfom calculations. The rnetabolic apparatus rnust be rneticulously and conectiy calibrated'06*"8~'"*'41v15g1"*'73with stnct adherence to protocol and rneasurernent conditions"8b136.For exarnple, the presence of anaesthetic gases renders the Haldane transformation void, therefore, to obtain an accurate rneasure of REE a washout period must be provided after the use of a general anaesthetic. A 24-hour period has been used in the dog by other investigator~'~'.The gas analysers require calibration with a known gas mixture that approximates the composition of the patient's expired air or encornpasses the range of expired The pneurnotach should be standardised with a calibration
~yringe'~~.An airtight rnask or canopy system is required to ensure no leakage of room air into the systern158.157 . Any individual with a physical, traumatic or surgical lesion resulting in air leaks should not be included in the study. Identifjdng and standardising appropnate clinical conditions for perforrning indirect calorimetry measurernents of REE will increase accuracy and precision106*118*126~'36~'37J72'73.
Indirect calorimetry is comrnonly used in humans to determine energy expenditure, to rnodify nutritional support, to establish a pattern of response dunng acute phase of illness, to evaluate the type and rate of substrate utilisation and to predict the success of weaning a patient from a ventilatorl 1*u*118*123*136~158v16060169.Single measurements of EE have recently been reported in postopeiative or traumatised dogs and in normal healthy dogs. The EE was measured during the early morning
(0500 - 08:OO) in a quiet, veterinary hospital room. The dogs were fasting for 12 hours, were receiving no medications that could alter the REE and had a rectal temperature within 1 to 2 ' C fmm normale5.The findings of this study demonstrated that the current estimate of EE, on average, using the predictive equation with an illness factor of 1.6 were significantly (p Thercfore, in critically il1 humans the only reliable method of deterrnining the 24-hour EE is to measure REE via indirect cal~rimet~~~~*'~~*'". Indirect calorimetry has been shown to provide reliable and repeatable measurements of EE in the healthy dogsl3'. Measuring REE throughout a 24-hour period may be the only way to provide a reliable means of assessing the 24-hour EE in il1 dogs. The measured REE may then be compared to the predictive equation currently published, The alternative method of assessing the 24-hour EE by indirect calorirnetry would provide veterinarians with a more reliable method of the determination of REE- The cornparison of the measured 24 EE with the predictive equation would provide insight into the current practice of assessing nutritional requirements in il1 dogs. 1.2.1 1 Summary: As nutritional support is an integral part of patient management during illness, undergraduate students and practising veterinarians are taught to include this as a routine component of the overall therapeutic plan for individual patients. However, the methods of nutritional support and calculation of EE described have not been validated in il1 or injured animals. The measurements of REE obtained by indirect dorimetry would be a valuable tool in determining nutrient requirements in hospitalised dogs. lndirect calorimetry has proved to be an accurate, reliable, non- invasive method of deterrnining the 24-hour EE in human patients. However, in veterinary medicine we are at the elementary stages of examining its use in clinical racticea.t30-1 32 33 1.2.1 2 Reference List: Chandra RK. lmmunodeficiency in undemutrition and ovemutrition. Nufr Rev 1981 ; 39:225-231. Meguid MM, Mughal MM, Meguid V, TehJT. Risk-Benefrt analysis of malnutrition and perioperative nutritional support: A review. Nufr Rep /nt 1987; 3~25-34. Pilbeam SPI Head A, Grossman GD, Sbnton GW, Heymsfield SB. Undemutrition and the respiratory system: Part 1. Metabolic and physiologic consequences. Respiratory Therapy 1983; Sept/Oct:65-69. Elwyn DH. Nutritional requirements of adult surgical patients. Crit Care Med 1980; 8:9-20. Kinney JM, Duke JH, Long CL, et al. Tissue fuel and weight loss after injury. J Clin Path 1970; 23:65-72. McClave SA, Spain DA. Indirect calorimetry should be used. NCP 1998; 131143-1 45. Armstrong PJ, Kirk CA, Gilson SD, Crane S.W. Enterai nutrition: Its importance in recovery. Hilk Science Diets Research Divison 1992; 2-36. Armstrong PJ, Saker KS, Davenport DJ, Remillard RL. Assisted feeding in hospitalized patients: Enteral and parenteral nutrition. 1n:fland MS, Thatcher CD, Remillard RL,et al ,eds. Small Animal Clinical Nutntion. Marceline,Missouri: Walsworth Publishing Company, 2000; 35 1-400. Donoghue S. Clinical nutrition :Nutritional support of hospitafized patients, in Small Animal Vetennary Clinics of Notth America 1 989; 475-495. Remillard RL, Thatcher CD. Critical care, in Veterianry Clinics of Nolth Amenca Small Animal Medicine 1989; 1287-1 306. Bistrian BR, Blackburn GL, Vitale J, et al. Prevalance of malnutrition in general medical patients. J Am Med Assoc 1976; 235:1567-1570. Cunningham KF, Aeberhardt LE, Wiggs BR, Phang PT. Appropriate interpretation of indirect calorimetry for detemining energy expenditure of ~atientsin intensive care units. Am J Sum 1994: 167547-550. Giner M, Laviano A, Meguid MM, Gleason JR. In 1995 a correlation between malnutrition and poor outcome in critically il1 patients still exits. Nutrition 1996; 12123-29. Klein SK, Kinney JM, Jeejeebhoy K, et al. Nutrition support in clinical practice: Review of published data and reccomendations for future research directions. J Parenter Enteral Nutr 1997; 21:t 33-1 56. Eichacker PO, Hoffman WD, Banks SM, Danner RL , Richmond S, Wilson L, MacVittie TJ, Natanson C.Serial metabolic cart and hemodynamic estimates of total body oxygen consumption in an awake canine model of human septic shock. Am Rev Respir Dis 1992; 145:AigO. Cerra FB. Hypermetabolism,organ failure, and metabolic support. Surgery 1987; 101:1-14. Shaw JHF, Wolfe RR. A conscious septic dog rnodel with hemodynamic responses similar to responses of humans. Surgery 1984; 95553-561. McClave SA, Snider HL. Understanding the metabolic response to critical illness: Factors that cause patients to deviate from the expected pattern of hypermetabolism. New Honi-Sci Pract 1994; 2:139-1 46. Mochizuki H, Trocki O, Dominioni L, Breckett KA, Joffe SN, Alexander JW. Mechanism of prevention of postburn hypermetabolism and catabolism by early enteral feeding. Ann Surg 1984; 200:297-308. Maguid MM, Collier MD, Howard LJ. Uncomplicated and stress starvation, in Surgical Clinics of North Amenca 1981 ; 532-543. VanLanschot JJB, Feenstra BWA, Vermeiji CG, et al. Calculation versus rneasurement of total energy expenditure. Crit Care Med 1986; l4:981-985. Roe CF, Goldberg MJ, Blair AB, Kinney JM. The influence on body temperature on early postoperative oxygen consumption. Surgery 1966; 60:85-92. Kinney JM, et al. Assessrnent of the energy metabolism in health and disease, in Proceedings. Report of the First Ross Conference on Medical Research 1980. Hasselgren PO, Jagenburg R, Karestrom L, et al. Change of protein metabolism in liver and skeletal muscle following trauma complications. J Traum 1984; 24:224-228. Barton RG, Cerra FB, The hypermetabolism-multiple organ failure syndrome. Chest 1989; 963 153. Dark DS, Pingleton SK. Nutrition and nuritional support in cntically il1 patients. J Intensive Care Med 1993; 8:16-33. Barton RG. Nutrition support in critical illness. NCP 1994; 9:127-139. Knreger KJ, DiPalma JA- The optimal diet for the critically il1 patient. Curr Opin Crit Care 1 997; 3:127-1 3 1. Minard G, Kudsk KA. ls eariy feeding beneficial? How eariy is early? New Horiz- Sci Pract Y 994; 2:156-163. Covelli HD, Black JW, Olsen MS, Beekman JF. Respiratory failure predicted by high carbo hydrate loads. Ann lntem Med 198 1; 95:579-58 1. Pilbeam SP, Head CA, Stanton GW, Grossman GD, Heymsfield SB. Undernutrition and the respiratory system. Part 2: Nutritional support. Respiratory Therapy 1983; Nov/Dec:72-78. Brooks MJ, Melnik G. The refeeding syndrome: An approach to understanding its complications and preventing its occurrence. Pharmocotherapy 1995; 15: 713. Sheldon GF, Peterson SR, Sariders R. Hepatic dysfunction dunng hyperalimentation. Arch Surg 1978; 113:504-508. Heymsfield SB, Hoff RD, Gray TF, Galloway J, Casper K. Heart diseases. 1n:Kinney JM, Jeejeebhoy KN, Hill GL,et al ,eds. Nufrifion and Metabolism in Patient Care. Philadelphia: WB Saunders, 1988; 477-509. Weissman C, Kemper M, Askanazi J, Hyrnan Al, Kinney JM. Resting metabofic rate of the critically il1 patient: Measured verses predicted. Anaesthesiology 1986; 64~673-679. Marino PL, Finnegan MJ. Nutritional support is not beneficial and can be harmful in criticaily il1 patients, in Crifical Care Clinicsr Controversies in Critical care Medicine 1986; 667-675. 37. Fettman MJ. Hyperrnetabolism in illness and injury. Comp Conf €duc Pract Vef 1992; 14:1284-129O. DeBaisse MA, Wilmore DW. What is optimal nutritional support? New Honi-Sci Pract 1994; 2:122. Guyton AC, Hall JE. Textbook of medicalphysiology. Philadelphia,London,Toronto :WB Saunders Company, 1996;3-1148. Wolfe BM, et al. Basic prinicples for surgical nutrition: Metabolic response to starvation , trauma and sepsis.eds. Nutrition in Clinical Surgefy. Baltimore: Willams and Wilkins, 1985; 14-23. Long CL. Energy balance and carbohydrate metabolism in infection and sepsis. Am J Clin Nutr 1977; 3O:l3Ol-l3lO. Wilmore DW. The Metabolic management of the critically ill. New York :Plenum, 1977. DeBruijne JJ, Altszuler N, Hampshire J, et al . Fat metabolism and plasma hormone levels in fasted dogs. Metabolis 1981; 30:190-194. Ballard FJ, Hanson RW, Kronfeld DS. Gluconeogenesis and lipogenesis in tissues from ruminant and nonruminant anirnals. Fed Proc 1969; 28:Sl8. Steeie R, Winkler B, Rathgeb 1, et al. Plasma glucose and free fatty acid metabolism in normal and long fasted dogs. Am J Physiol 1968; 21 4:3l3. Brady LJ, Armstrong MK, Muinin' KL. Influence of prolonged fasting in the dog on glucose turnover and blood metabolites. J Nufr 1977; 1O7:lO53. Balasse EO. Evidence for an effect of insulin on the peripheral utilisation of ketone bodies in dogs. J Clin lnvest 1971; 50:8Ol-813. Pi-Sunyer FX. Starvation-induced ketosis:Reduction in dogs enriched with odd- carbon fatty acids. P Soc Exp Bi01 Med 1974; 145786. Walters JM, Bessey PQ, Dinarello CA, Wolff SM, Wilmore DW. Both inflamrnatory and endocrine mediators stimulate host responses to sepsis. Arch Surg 1986; 121:179-l9O. The lams Company-Foodfor thought: Nutritional management of stress:Critical care. 1999. Kinney JM. Clinical biochemistry: Implications for nutritional support. J Parenter Entera1 Nutr 1990; 14:148S-l56S. Burkholder MJ, Swecker WSJ. Nutritional influences on immunity. Sernin Vef Med Surg ( Srnall Anim) 1 990; 5:1 54-1 66. Bursztein S, Elwyn DH, Askanazi J, Kinney JM . Theoretical framework of indirect calorimetry and energy balance-eds. Energy Metabolism, Indirect Calorimefry and Nurition. Baltimore, Maryland: Williams & Wilkens, 1989; 27-81. Li M, Specian RD, Berg RD, Deitch EA. Effects of protein malnutrition and endotoxin on the intestinal mucosal barrier to the translocation of indigenous flora in mice. J Parenter Enteral Nutr 1989; 13:572-578. Souba WW. Cytokine control of nutrition and metabolisrn in critical iilness. Curr Prob Surg 1994; 3 1:582-642. Michie HR, Spriggs DR, Manogue KR, et al. Tumor necrosis factor and endotoxin induce similar metabolic responses in human beings. Surgery 1988; 104~280-286. Nelson KM, Long CL. Physiologie basis for nutrition in sepsis. NCP 1989; 4:6- 15. Shaw JHF, Wildbore M,Wolfe RR. Whole body protein kinetics in severely septic patients: The response to glucose infusion and total parenteral nutrition. Ann Surg 1987; 205:288-294. Cerra FB, Siegel JH, Coleman B, et al. Septic autocannibalisrn: a failure of exogenous nutritional support. Ann Surg 1980; 192570-580. Kala J, Ruokonen E, Webster NR, Nielsen MS, Vundelinckx G, Zandstra DF. Increased mortality associated with growth hormone treatment in critically il1 adults. N Engl J Med 1999; 341 :785-792. Fried RC, Bailey PM, Mullen JL, Stein TP, Crosby LO, Buzby GP. Aiterations in exogenous substrate rnetabolisrn in sepsis. Arch Surg 1986; 121: 173-1 78. Wolfe RR, Durkot MJ, Wolfe MH. Effect of thermal injury on energy metabolism, su bstrate kinetics, and hormonal concentrations. Circulatory Shock 1982; 9:383-394. Kronfeld DS. Protein and energy estimates for hospitalized dogs and cats, in Proceedings. Purina Int Nutr Sym in assoc East Vet Conf Orlando 1991; 5-1 1. Long CL, Schaffel N, Geiger JW, Schiller WR, Blakemore WS. Metabolic response to injury and illness: Estimation of energy and protein needs Rom indirect calorimetry and nitrogen balance. J Parenter Enteral Nufr 1979; 3:452- 456. Walton RS, Wingfield WE, Ogilvie GK, Fettrnan MJ, Matteson VL. Energy expenditure in 104 postoperative and traumatically injured dogs with indirect calorimetry. J Vet Emerg Crif Care 1998; 6:71-79. ASPEN Board of Directors. Guidelines for the use of parenteral and entera1 nutrition in adults and paediatric patients. J Parenter Enferal Nutr 1993; 17:l SA-51 SA. Hand MS, Crane S.W., Buffington CA. Surgical nutrition. 1n:Betts CW, Crane SW ,eds. Manuel of Small Animal Surgical Therapeufics. New York: Churchill Livingstone, 1986; 91-1 15. Donoghue S. A quantitative summary of nutritional support services in a veterinary teaching hospital. Comell Vet 1991; 81 : 109-1 28. Garre MA, Boles JM, Youinou PY. Current concepts in immune derangement due to undernutrition. J Parenter Enferal Nutr 1986; 1 1:309-313. Law D, Stanley J, Durïck MD, Abdou NI. lmmunocompetence of patients with protein-calorie malnutrition. Ann lntern Med 1973; 79545-550. Delves PJ, Roitt IM. Advances in immuno1ogy:The immune system ( first of two parts). N Engl J Med 2000; 343:37-49. Delves PJ, Roitt IM. Advances in immuno1ogy:The immune system ( second of two parts). N Engl J Med 2000; 343:108-ll7. Temple WJ, Voitu A, Snelling PT. Effect of nutrition, diet and suture material on long term wound healing. Ann Surg 1975; 182:93-97. Law NW, Ellis H. The effect of parenteral nutrition on the healing of abdominal wall wounds and colonic anastomoses in protein-malnounshed rats. Surgery 1990; 107:449454. Singer AJ, Clark MF. Mechanisms of disease: Cutaneous wound healing. N Engl J Med 1999; 341:738-746. Abel RM, Grimes JB, et al. Adverse hemodynarnic and ultrastnrctural changes in dog hearts subject to protein-caloric malnutrition. Am J Head 1979; 97:733- 744. Heyrnsfield SB, Bethe1 RA, Ansley JD, Gibbs DM , Felner JM, Nutter DO. Cardiac abnormalities in cachectic patients before and during nutritional repletion. Am J Hearf 1978; 95:584-594. Askanazi J, Weissman C, Rosenbaum-S H. Nutrition and the respiratory system. Crif Care Med 1982; 1O:l63-172. Doekel RC, Zwillich CW, Scoggin CH, Kryger M, Weil JB. Clincial serni- starvation: Depression of hypoxic ventilatory response. N Engi J Med 1976; 2951358-361. Rosenbaum SHI Askanazi JI Hyman Al, et al-Respiratory patterns in profound nutritional depletion. Anaesthesiology i 978;51 :S366. Arora NE, Rochester DF. Respiratory muscle strength and maximal voluntary ventilation in undemourished patients. Am Rev Respir Dis 1982; IZ6:508. Zaloga GP, Roberts PR. Early enteral feeding improves outcome. Yearbook of Intensive Care and Emergency Medicine. 1 997; 70 1-7 14. Moore FA, Moore EE Trauma. 1n:Zaloga GP ,eds. Nutrifion in Critical Care. St. Louis,Missouri: Mosby-Year book lnc , 1994; 571-586. Kemper M, Weissman CI Hyman Al. Caloric requirements and supply in critically il1 surgical patients. Crif Care Med 1992; 20344-348. Park GR. Pharmacology, metabolism, and nutrition. CUIT Opin Crit Care 1997; 3~253-254. Walter-Sack 1, Klotz U. Influence of Diet and nutritional status on drug metabolism. Clin Phamocokinet 1996; 31 :47-64. Bursztein S, Elwyn DH, Askanazi JI Kinney JM , Kvetan V, Rothkopf MM, Weissman C. Indirect calorimetry: History and overview. 1n:Bursztein S, Elwyn DH, As kanazi J ,et al ,eds. Energy Mefabolism, lndirect Calonmetry, and Nutrition. Baltimore, Maryland: Willams & Wilkins, 1989; 2-23. Weissman CIHyman Al. Nutritional care of the critically il1 patient with respiratory failure. Crit Cam Med 1987; 3:l85-203. Lowry SF, Brennan MF. Abnomal liver function during parenteral nutrition: Relation to infusion excess. J Surg Res 1979; 26:300-307. Mashirna Y. Effect of calorie overioad on puppy livers during parenteral nutrition. J Parenter Enteral Nufr 1979; 3:139-145. Erickson KL. Dietary fat modulation of immune response. Inf J lmmunopharmaco 1986; 8:529-543. Johnson PV. Dietary fat, eicosanoids, and irnmunity. Adv tipid Res 1985; 21:lO3-141. Hwang D. Essential fatty acids and immune response. FASEB J 1986; 3:2052- 2061. Lee TH, Hoover RL, Willarns JD, et al. Effect of dietary enrichment with eicosapentaenoic and docosahexaenoic acids on in vitro neutrophil and monocyte leukotriene generation and neutrophil function. N Engl J Med 1985; 31211 21 7-1 224. Ciander B, Clandinin MT, Hosokawa T, et al. Dietary fat alters the fatty acid composition of lymphocyte membranes and the rate at which suppressor activity is lost. Immun01 Leff 1983; 6:33l-337. Wolfe BM, Chock E. Energy sources, stores, and hormonal controls. Surg Clin N Am 1981; 61 :SO9-518. Feurer 1, Crosby LO, Mullen JL. Measured and predicted resting energy expenditure in clinically stable patients. Clin Nutr 1984; 3:27-34. Mann S, Westenskow DR, Houtchens BA. Measured and predicted caloric expenditure in the acutely ill. Crit Céire Med 1985; 13: 173-177. Roza AM, Shizgal HM. The Hams Benedict equation reevaluated: Resting energy requirements and the body cell mass. Am J Clin Nutr 1984; 40:168-182. 100. Damask MC, Schwartz Y, Weissrnan C. Energy measurements and requirements of critically il1 patients. Crit Care Clin 1987; 3:71-96. 101. EIwyn DH, Kinney JM, Askanazi J. Energy expenditure in surgical patients. Surg Clin NAm 1981; 61:545-555. 102. Hunter DC, Jaksic T, Lewis D, Benotti PN, Blackburn GL, Bistrian BR. Resting energy expenditure in the critically il[: Estimations versus measurement. Br J Surg 1988; 75:875-878. 103- Burkholder MJ. Metabolic requirements of sick dogs and cab. J Am Vet Med ASSOC 1995; 2061614-618. 104. Kleiber M. The tire of iife: An introduction to animal energetics- Huntington, New York :Kneger, 1975. 105. Brody SIProcter RC, Ashworth US.Basal metabolism,endogenous metabolism and neutral sulphur excretions as a function of body weight. 1934;Research Bulletin. 106. McClave SA, Snider HL, Greene L, et al. Effective utlization of indirect calorimetry during critical care. Intensive Care Wodd 1992; 9:194-200. 107. Delany JP, Lovejoy JC. Energy expenditure. Endocrin Metab Clin 1996; 25:831. 108. Boothby WM, Sandiford 1. Basal metabolism. Physiol Rev 1924; 4:69-151. 109. Benedict FG. The measurement and significance of basal metabolism.eds. Lectures on nutrition:A senés of lectures given a the Mayo foundafion and the University of Wisconsin. Philadelphia: WB Saunders, 1925; 17-58. 110. Boothby WM, Berkson JI Dunn HL. Studies of the energy expenditure of noma1 individuals: A standard for basal metabolism with a nomogram for clinicat application. Am J Physioi 1936; 3:468-483. 111. Boothby WM, Sandiford 1. Summary of the basal metabolism data on 8614 subjects with especial reference to the normal standards for the estimation of basal metabolic rate. J Bi01 Chem 1922; 54:783-803. 112. Harris JA, Benedict FG. Standard basal metabolism constants for physiologists and clinicians, eds. A Biometnc studies of basal metabolism in man. Phifadelphia: JB Lippincott, 1919; 223-250. 113. Gump FE, Martin Pl Kinney JM. Oxygen consumption and caloric expenditure in surgical patisnts. Surg Gynecol Obstetr 1973; 137:499-513. 1 14. Blaxter KL. The minimal metabolism. In:8laxter KL ,eds. Energy metabolism in animals and man. Cambridge: Cam bridge University Press, 1989; 120-146. 1 15. Crist KA, Romos DR. Evidence for cold-induced but not for diet-induced thennogenesis in adult dogs. J Nutr 1987; 1 li:lZ8O-l286. 116. Kienzle E, Rainbird A. Maintenance energy requirements of dogs: What is the correct value for the calculation of metabolic body weight in dogs? J Nutr 1991 ; 121:S39S40. 117. Manner K. Energy requirement for maintenance of adutt dogs. J Nufr 1991; 121:S37S38. 12 8. McClave SA, Snider HL, Use of indirect calorimetry in clinical nutrition. NCP 1992; 7:207-221. 119. Altman PL, Dittmer DS. Biology data book. In:Bethesda,MD eds :Federation of American Societies for Experirnental Biology, 1973. 120. Francisco R, Aquin L, Seades JM, Banchero N. Oxygen cost of breathing in dogs. Respiration 1978; 35186-1 9 1. 121. Diamond P, Brondel L, LeBlanc J. Palatability and postprandial thermogenesis in dogs. Am J Physiol 1985; 248:E75-E79. 122. LeBlanc J, Diamond P. Effect of meal size and frequency on postprandial thermogenesis in dogs. Am J Physiol 1986; 250:El44-El47. 123. Chiolero Rt, Bracco O, Revelly JP. Does indirect calorimetry reflect energy expenditure in the criticaliy il1 patient. in:Berlin, Wilmore DW, Carpentier YA ,eds. Metabolic Support of the Critically Ill Patient, 1993; 95-1 18. 124. Swinamer DL, Phang PT, Jones RL, Grace M, King EG. Twenty-four hour energy expenditure in criticafly il1 patients. Crif Care Med 1987; 15:637-643. 125. Askanazi J, Carpentier YA, Elwyn DH, Nordenstrom J, et al. Influence of total parenteral nutrition on fuel utilisation in injury and sepsis. Ann Surg 1980; 191 ~40-46. 126. Weissman C, Kemper M, Damask MC, et al. Effect of routine intensive care interactions on metabolic rate. Chest 1984; 86:815-818. 127. Leff ML, Hill JO, Yates AA, Cotsonis GA, Heymsfield SB. Resting rnetabolic rate: Measurernent reliability. J Parenter Entera1 Nutr 1987; 11 :354-359. Aschoff J, Pohl H. Rhythmic variations in energy metabolism. Fed Proc 1970; 2911541 -1 552. VanLanschot JJB, Feenstra BWA, Vermeiji CG, et al. Accuracy of intermittent metabolic gas exchange recordings extrapolated for diurnal variation. Crif Care Med 1988; l6:737. Walters LM, Ogilvie GK, Salman MD, Fettman MJ, Joy J, Hand MS, Wheeler SL. Repeatability of energy expenditure measurement in clinically normal dogs by use of indirect calorimetry. Am J Vet Res 1993; 54:1881-1885. Ogilvie GK, Salrnan MD, Kesel L, Fettman MJ. Effect of anaesthesia and surgery on energy expenditure detemined by indirect calorimetry in dogs with malignant and non-rnalignant conditions. Am J Vet Res 1996; 57:13Zl-l325. Greco DS, Rosychuk RAW, Ogilvie GK, Harpold LM, Van Liew CH. The effects of levothyroxine treatment on resting energy expenditure of hypothyroid dogs. J Vet lntem Med 1998; 12:7-10. Peters J, Arndt AO. Letter to the editor: Fluctuations in resting energy expenditure. Cnf Care Med 1984; 926. Elliott DA, King LG, Zerbe CA. Thyroid hormone concentrations in cntically il1 canine intensive care patients. J Vef Emerg Crit Care 1995; 5:V-23. Damask MC, Weissman C, Askanazi J. Letter to the editor a reply: To fluctuations in fasting energy expenditure. Crif Care Med 1984; 926-926. Feurer 1, MuIlen JL. Bedside measurement of resting energy expenditure and respiratory quotient via indirect calorimetry. NCP 1986; 1 :43-46. Makk LJK, McClave SA, Creech PW, Johnson DR, Short AF, Whitlow NL, Priddy FS, Sexton LK, Simpson P. Clinicat application of the metabolic cart to the delïvery of total parenteral nutrition. Cnt Care Med 1990; 18:1320-1327. Borgardis C, Lilliojia S, Ravussin E, et al. Familial dependence of resting metabolic rate. N Engl J Med 1986; 315 :96-100. Swinamer DL, et al. Effect of routine administration of analgesia on energy expenditure in cntically il1 patient. Chest 1988; 92:4-10. Damask MC, Askanazi J, Weissrnan Cl Elwyn DH, Kinney JM. Artifacts in the measurernent of resting energy expenditure: CCnt Care ~ed1983; 1 1 :750-752. 141. Bursztein S, Elwyn DH, Askanazi J, Kinney JM . Methods of measurement and interpretation of indirect calorimetery.eds, Energy Metabofism, Indirect Calorimetry and Nutrition, Baltimore, Maryland: Williams & Wiikens, 1989; P 73- 208. 142. Heusner AA. Body mass,maintenance and basal metabolism in dogs. J Nufr 1991; 121:~8-~17. 143. Lusk G, DuBois EF. On the constancy of the basal metabolism. J Physiof 1924; 59~213-216. 144. Abrams JT. Diets, culture media and food supplement. The nutrition of the dog, in CRC Handbook Series in Nutrition and Food 1977; 3-28. 145. Burger IH, Johnson JU. Dogs large and small: the allometry of energy requirements within a single species. J Nutr 1991; 121:Si 8621. 146. Calder WA3. Scaling of physiological processes in homeotherrnic animals. Ann Rev Physiol 198 1 ; 43:3Ol-322. 147. Galvao PE. Heat production in relation to body weight and body surface.lnapplicability of the surface law on dogs of the tropical zone. Am J Physiol 1946; l48:478-490. 148. Daly JM, Heymsfield SB, Head CA, Hawey LP, Katzeff HL, Grossman GD. Human energy requirements: overestimation by widely used prediction equations. Am J Clin Nutr 1985; 421170-1 174. 149. Ireton-Jones CS, Jones JD. Should predictive equations or indirect calorimetry be used to design nutrition support regimens? NCP 1998; l3:l4l-l4S. 150. Hansell DT, Davies JWL, Gisbey EM, Gilmour WH, Burns HJG. Comparison of various predictive equations for the estimation of resting energy expenditure. Clin Nutr 1989; 8:299-305. 151. Cortes V, Nelson LD. Errors in estimatlng energy expenditure in critically il1 surgical patients. Arch Surg 1989; 124:287-290. 152. Bursztein S, Saphar P, Singer Pl Elwyn DH. A mathematical analysis of indirect calorimetry measurements in acutely il1 patients. Am J CIin Nutr 1984; 50:227- 230. 153. Baker JP, Eesky AS, Stewart S, et al. Randomization trial of total parenteral nutrition in critically il1 patients: Metabolic effects of varying glucose-lipid ratios as energy source. Gasfroenferology 1984; 87:53-59. 154. Lavoisier AL. Elements of chemisfry in a new systematic order, Edinburgh :Creech, 1793. 155. Foregger R. Lavoisier's Research on Carbon Dioxide and its Control. Anaesthesiology 1962; 23:662-669. 156. Brandi LS, Bertolini R, Calafa M. Indirect calorimetry in critically il1 patients: Clinical applications and practical advice. NutMon 1997; 13:349-358. 157. Ferrannini E. The theoretical basis of indirect calorimetry: A review. Metabolis 1988; 37~287-301. 158. Kinney JM. Energy metabolism. 1n:Fischer E ,eds. Surgical Nutrifion. Boston: Little Brown and Company, 1983. 159. Weir JB. New methods for calculating metabolic rate with special reference to protein metabolism. J Appl Physiol 1949; 109:l-9. 160. Kinney JM. Indirect calorimetry in malnutrition: Nutritional assessrnent or therapeutic reference? J Parenter Enteral Nutr 1987; 11 :90S-94s. 161. Klotz IM. Introduction to biomolecular energefics:Including ligand-recepfor interactions. Odando,San Diego,New York :Academic Press Inc., 1986; 162. Lusk G. The elements of the science of nufntion. Philadelphia :Saunders, 1931. 163. Rothwell NJ, Stock MJ. Regulation of energy balance. Ann Rev Nutr 1981; 1:235. 164. Frascarolo P, Schutz Y, Jequier E. Decreased thermal conductance during the luteal phase of the menstrual cycle in women. J Appl Physiol 1990; 69:2029- 2033. 165. Operator's Manual, Aerosport Inc. TEEM 100 Metabolic analysis system total energy expenditure measurement. TEEM 700 Metabolic Analysis System Operafor's Manual 1993; 147. 166. Elia M, Livesey G. Theory and vafidm of indirect calorimetry during net lipid synthesis. Am J Clin Nutr 1988; 47591-607. 167. Weissman CsOxygen transportafion and ufilisation, Fullerton Ca :The Society of Critical Care Medicine, 1987;25-64. 168. Head CA, McManus CB, Seitz S, Grossman GD, Stanton GW, Heymsfield SB. A simple and accurate indirect calorimetry system for assessment of resting energy expenditure. J Parenfer Enteral Nufr 1984; 8:45-48. 169. Branson RD. The measurement of energy expenditure: Instrumentation, practical considerations and clinical applications. Respiratory Care 1990; 351640-659. 170. Makita K, Nunn JF, Roysten B. Evaluation of metabolic measuring instruments for the use in critically il1 patients. Crif Care Med 1990; 18 :638. 171. Ultman JS, Bursztein S. Analysis of error in the determination of respiratory gas exchange at varying F102.J Appl Physiol 1981; 50:210-216. 172. Mullen JL. Indirect calorimetry in critical care. P Nufr Soc 1991; 50:234-244. 173. Zavala DC. Evaluating nutritional needs.eds. Nutrifional Assessrnent in Crifical care: A Training Handbook. lowa City: University of lowa press, 1989; 41 -59. 174. Matarese LE. Indirect calorimetry: Technical aspects. J Am Diet Assoc 1997; 97:Sl54Si 60. 175. Choliero R, Flatt JP, ReveIly JP, et al. Catecholamines on 02 consumption and oxygen delivery in critically il[ patients. Chest 1991; 100:1676-1684. 176. Boyd O, Grounds SM, Bennett D. The dependence of oxygen consumption on 02 delivery in critically il1 postoperative patients is mimicked by variations in sedation. Chest 1992; 101:A61 9-1 624. Cha~ter2: Evaluation of the Accuracy and Reliabiiity of Indirect Calorimetry for the Measurement of Resting Energy Expenditure in Heaithy Dogs. 2.0 Abstract Objective: The objective of this study was to assess the accuracy and reliability of open-flow indirect calorimetry in healthy dogs. First, two methods of indirect calorimetry, the open-flow indirect calorimeter and the closed-circuit spirometer, for the measurement of oxygen consumption on a per kilogram basis (V02/kg) were assessed for the level of chnical agreement. Second, the reliability of the open-flow indirect calorimetry was assessed by evaluating the reproducibility of the RE€ and the V02 measurements in healthy dogs over a two-day penod. Animals: A total of 13 healthy dogs were enrolled. Six were mixed breed research dogs ranging in weight from ? 9 to 30 kg and age from 1 to 3 years enrolled in the first phase of the study; seven were healthy staff-owned dogs ranging in weight from 21.1 to 33.3 kg and age from 2 to 10 years enrolled in the second phase. Procedure: In the first phase, the six research dogs' measurements of VOz /kg were perfomed while intubated and rnaintained under propofol general anaesthesia. Once a stable plane of anaesthesia was achieved, five one-minute V02/kg measurements were recorded by the open-flow indirect calonmeter (pre-IC) followed by a four one- minute V02/kg measurements by the closed-circuit spirometer (SP) and then the open-flow indirect calonmeter (post-IC) measurements were repeated. The entire cycle was then repeated. In the second phase, four series of 16 one-minute readings of REE and VOawere perforrned by the open-flow indirect calorimeter on each of the seven staff owned dogs at the same time of day for two consecutive days. A series consisted of a 1O-minute adaptation period to the collection system and six one- minute sampling periods. The mean VOz and carbon dioxide production (VC02) of the six one-minute sampling periods were used to calculate the REE by using the Manuscriptwritten in the format of aie American Journal ofveterinary Research abbreviated Weir for~nula.'~The same two individuals perfomed al1 open-flow indirect calorimetry measurements in a themoneutml environment (20 to 22" C). Results: In the first phase, the mean values for the V02/kg measurernents between the pre-IC and post-IC, were significantly different (p=0.018) with the pre- IC readings being consistently greater than the post-IC readings. An analysis of agreement was performed and revealed that the pre and post-IC rnethods for the determination of V02/kg had an acceptable level of clinical agreement with the closed-circuit spirometer. In the second phase, the interclass correlation coefficient (CC)for al1 series, the REE was 0.779 and the VOz was 0.786. A more reliable measure of the REE (ICC=0.904) and VOz(ICC=0.894) was observed if the first series was discounted. The most reliable and least variable measures of REE and V02 were obtained if the first two series were discounted. There were two significant ho-way interactions (dog'series, dog'day) and one three-way interaction (dog*day*series) observed. lndicating that the differences between two dogs or within the same dog is not similar over the two days studied and the magnitude of this difference depends upon which of the four series measured was used in the analysis. Conclusions and Clinicat relevance: Despite the inherent variability in the rneasurement of REE, the reliability of serial open-flow indirect calorimetric metabolic measurernents is good. However, measurements of REE and V02 are only reliable when a IOminute adaptation period to the collection system is perfomed and the first series, of a series of measurements, is discounted. The variance observed in the most reliable series (three and four) was 9500.96. An individual reading in either of the series three or four maybe, on average, 7.27 kcal/kg/day above or below the mean. Open-flow indirect calorimetry is a clinically acceptable method for the determination of V02/kg and therefore REE in healthy dogs and will provide a more flexible and feasible method for the determination of the 24-hour energy expenditure than closed-circuit spirometry in a clinical setting. Manuscript written in the fonnat of the American Journal dvetennary Research 49 < 2.1 Introduction: Malnutrition or overfeeding rnay adversely affect clinical outcome in critically il1 humans? Therefore, a precise assessment of the resting energy expenditure (REE kcallday) has becorne an mtegral part of nutritional support in h~rnans?~A reliable approach to assessing the daily energy expenditure is to measure the REE under a standardised set of ~onditions.'~~.~~~*~The standard basal conditions in humans for the measurement of basal energy expenditure are as follows: first thing in the morning, after an ovemight sleep, fasted for at least 1O hours, with a normal body temperature, in a supine position and in a thermoneutral environment with no physiologic or psychological stress- As it is impractical to achieve basal conditions for the measurement of EE in hospitalised humans, energy expenditure is measured under resting conditions. The "resting* conditions under which REE is measured are less strictly defined and may include the following situations: periods of diet-induced theAogenesis, physiologic or psychological stress. and variations in body and environmental temperature.1° It is important that a clear definition of the conditions under which the rnetabolic rate was measured be provided in the experimental protocol to facilitate comparisons amongst researchers? Indirect calorimetry, which measures gas exchange [oxygen wnsumption 0(02)and carbon dioxide production (VC02)]at the level of the lungs, is considered the current standard for measuring the REE in hospitalised humans?' These gas exchange measurements are converted to kilocalories per day by the application of the abbreviated Weir formula or by calculating the caloric equivalent of the litres of oxygen consumed and carbon dioxide produced with the use of standardised table^.'^"^ There is extensive information in the human medical literature on the reliabi~itf*'~"~,accura~y'"~ and clinical application of indirect ca~orirnetry~~~~; however, this technique has just begun to be examined in veterinary rnedicii~e~~~~. Manuscript Wtten in the format of ttie American Journal of Veteflnary Research This study was designed as two separate, but retated phases. The objective of the studies reported was to assess the accuracy and reliability of an open-flow indirect calorimeter in healthy dogs. First, the open-flow calorimeter was assessed by comparing it to a traditional clinical standard, closed-circuit spirometer, for the measurement of oxygen consumption on a per kilogram basis 0(02/kg). Second, the reliability of open-flow indirect calorimetry was assessed by evaluating the reproducibifity of RE€ and V02 rneasurements in healthy dogs, over a two-day period. The V02 /kg determined by open-flow indirect calorimetry was hypothesised to agree with the V02/kg measured by closed-circuit spirometry. Furthermore, the measures of REE and V02as determined by open-flow indirect calorimetry was hypothesised to demonstrate a clinically acceptable level of reliability in healthy dogs. 2.2 Material and Methods: Phase I- Six healthy research dogs were used to assess agreement of the measurement of V02/kg (mI/min/kg) between the two indirect calorimetry methods; the open-flow indirect calorimete? and the closed circuit spiromete?. All dogs were determined to be clinicalIy normal on the basis of a physical examination, hemogram, and biochemical profile. The dogs ranged in weight from 19 to 30 kg and in age from two to three years. Three were intact females and the remainder intact males. All were in good physical condition and none were oveweight. The RE€ was measured in fasted dogs in a therrnoneutral environment (22" C). A thermoneutral environment is defined as the ambient temperature, above or below, which REE of a nude subject begins to increase or decrease to maintain body temperature and normal heat balance? The anaesthetic protocol consisted of premedication with acepromazinaC (0.02 mglkg) and butorphenold (0.2 mglkg) administered intramuscularly. The dogs were induced with propofoIe(4 mg/kg) and general anaesthesia was maintained with Manuscript Menin the format of the Amencan Journal of Veterinary Research a constant rate infusion (0.2 mglkglhr or to effect) of propofole. Orotracheal intubation was performed on al1 dogs which were maintained on room air, Body temperature was maintained between 37.2 to 38 degrees Celsius with a circulating water blanketf and blankets. Body temperature, blood pressure, blood oxygen saturation and inspired oxygen content were monitored throughout the experiment at 15-minute intervals. Once a stable plane of anaesthesia was obtained the following measurements were performed: (1) A 5-minute V02/ kg was determined by the open- flow indirect calorimeter (pre-IC). (2) The dog's endotracheal tube was then attached to the closed circuit spirometer with a kymograph for a 4-minute period. A pressure test was performed prior to each spirometry test to ensure no air leakage was present within the system. One hundred percent oxygen was added to the seven litre bellows such that it occupied one ffih of the volume. This increased the inspired oxygen content from 21 % to approximately 25% and ensured that the dogs did not become hypoxemic when breathing the gas mixture frorn the spirometer. (3) A second 5- minute V02/kg measurement was then performed on the open-flow indirect calorimeter (post-IC)- The entire procedure was then repeated twice more for a total of four 5-minute open-flow indirect caiorimetry readings (two pre-IC and two-post-IC) and two 4-minute closed-circuit indirect calon'meter, water-filled spirometry readings of VOdkg each interposed between the pre and post-IC tests. A 10-minute penod was allowed between each procedure. The ambient temperature and barometric pressure were recorded on each day. The spirometer system was flushed with room air and 100% oxygen added pnor to each reading. Mettic graph paper (millimetre gradations) was applied to the rotating drum on the spirometer. The speed of the drum rotation was 32-mm/ minute. A 1-mm in height was equivalent to 20.8 mi, which was determined by injecting a known quantity of air through the system using a volume calibration syringe. The mean value for V02was obtained by dividing the change in the volume by the change in time on the kymograph record. The gas Manuscript written in the format of the Amencan Journal ofveterinary Research volumes recorded were then corrected to standard conditions namely 0° C for temperature. 760 mmHg for barometric pressure and as dry gases (STPD). Phase II- The reliability study of REE and V02was performed on seven staff-owned animals. The dogs ranged in weight from 21 -1 to 33.7 kg and in age from 2 to 9 years; four were spayed femaIes and three were neutered males. Ail were within IO% of ideal body weight as determined by physical examination, with no previous medical history and no known exposure to anaesthesia or exogenous corticosteriods within the previous rnonth. All dogs were determined to be clinically normal on the basis of a physical examination, hemogram, biochemical profile and free serum thyroxine concentration. These dogs received a similar dietg and were on a similar feeding schedule prior to and during the study. The duration of time between a meaI and the serial REE measurements was consistent over the two days but the initial rneasurement times differed in each dog (e-g. dog 1 ate at 08:OO and the REE determinations were performed at 09:OO. Dog 2 ate at 08:OO and the REE determinations were at 11:00). The REE measures were perfomed in a thermoneutral environment with the dogs in lateral recumbency after an adaptation period to the surroundings. A series of four 16-minute readings were performed on each dog. at the same tirne of day, over two consecutive days. The 16-minute reading, which comprised a series, was divided into a 10-minute adaptation period to the facemask and collection system and six one-minute data collection periods. The mean Von and VC02 were calculated from the six one-minute readings. The REE was then calculated from the mean V02and VC02 rneasurements by using the abbreviated Weir equation. Protocols were following in accordance with the guidelines set by the Canadian Council on Animal Care ". Open-tïow Indirect Calorimetry- Manuscript Menin the fmatof the Arnerican Journal 3fVeterinary Research The same two individuals obtained ail measurements of REE and V02, using a portable open-flow indirect calorimetry system. The instrument" used measures the partial pressure of oxygen and carbon dioxide in expired air by a galvanic fuel cell and a nondispersive infrared analyser, respectively. Expiratory ventilation volume is measured with a flat plate orifice pneumotach. The flow of gas through the sharp edged, thin plate pneumotach orifice produces a pressure drop proportional to the square root of the fiow. This flow signal is then integrated to obtain the expired air volume. lndependent instruments within the system measure temperature, barometn'c pressure and time. A sample of the expired air, the volume of which is proportional to the total expired volume for that breath, is admitted to the mixing chamber via a high frequency sampling valve. A fixed rate of proportional sampling, known as a pulse, is drawn into the mixing chamber. For each pulse drawn into the mixing chamber, a pulse of identical volume is emitted from the mixing chamber to the oxygen and carbon dioxide sensors. Over a fixed time period, electronic variable sampling alIows the pulse trains to be reduced to a constant volume, resulting in similar equilibration times at varying expired flow rates. This system of expired gas sampling is a modification of the breath-by-breath technique. The microprocessor calculates V02, VC02 and respiratory quotient (RQ) using standard equations.28 Before use, the gas analysers were calibrated to a standard gas of known composition (18.0 % O2and 2.0% CO2), which was similar to the expired air composition of dogs, on a daiIy basis and calibrated on room air prior to each test. The pneumotach was calibrated with a 1-L calibration syringeh prior to each test. Measurements of V02and VC02were calculated every 20 seconds and an average of these readings were printed at one- minute intervals. Burstein et alz0demonstrated in critically il1 humans that there is on average a 2 % error in the REE if protein metabolism is not taken into effect. Therefore, the REE was calculated by using the abbreviated Weir forrnu~a~~~~:REE (kcallday) = [3.94 (W2) + 1.1 (VC02)] X 1440 where V02= O2consumption (Umin) Manuscriptwritten in ihe fmatof the Ameflcan Joumat ofveterinary Research and VC02 = CO2production in (Umin), in which the nitrogen urine content is discounted. Statïstical Analysis: In order to assess the effect of time on the pre and post-IC readings in the agreement study a repeated rneasure ANOVA was performed, which assumed a constant correlation over time within an animal. A paired t-test was also performed on the two spirometry readings. A limit of agreement analysis, as described by Bland and ~itrnan~',was used to determine the level of agreement between the pre-IC. post-IC and the spirometry measurements of VOz/ kg. Strictly for companson with other studies, a Pearson's correlation coefficient was calculated for the pre and post open-flow indirect calorimeter V02/kg rneasurements to the closed-circuit spirometer VOdkg rneasurements. The reliability of the REE and V02measurements was estimated by calculating an intraclass correlation coefficient (ICC). The ICC was estimated frorn an analysis of variance, which exarnined the different variance components: dog, day, series and their possible interactions: dog'day, dog*series, dog*day'series, in a nested-split plot mode1 and accounted for the auto-correlation in the data by fitting a time series. The CC is a statistical tool used to evaluate the reliabiiity of measurements by estirnating the degree of change from one measurement to the next for a particular test?' An ICC equal to 1.O indicates that there is no within subject variance associated with a test. Conversely an ICC of zero indicates that there is no reliability associated with a test. Any ICC greater than 0.8 is considered to represent good reliability in a test, however an ICC greater than 0.7 is ;\cceptab~e.'~'~*'~An ICC was also determined for a subset of the series measurements in the same fashion. The means were reported as best Iinear unbiased estirnates of the REE and V02 (BLUPS). The BLUPS are best estirnates of what the actual REE or V02will be for an individual when al1 the different variance components: dog, day, series, dog'day, Manuscript wntten in the format of the Amencan Journal of Veterinary Research dog*series, dog*day*series, have been taken into account. They are interpreted in a similar manner to the means derived from a fixed effects model, however, they are the appropriate estimators of means in a random effects model.J2 Probability values of an alpha less than 0.05 were considered statistically significant. Al1 data analyses were performed using a commercial statistical software packagei. 2.3 Results: Agreement Study: A steady plane of anaesthesia was achieved in the six research dogs and were maintained throughout the expenment. The room (21 to 22" C) and body temperature (37.2 to 3a°C), blood pressure (mean arterial pressure between 60 to 70 mmHg) and blood oxygen saturation (greater than 94%) were maintained within acceptable Iimits throughout the procedure. The repeated measures ANOVA suggested that time had an effect on the indirect calorimetry measurements (p=0.08). Therefore, a contrast t-test was performed to identify a possible time effect on the pre-IC readings, which were then compared to the post-lC readings. There was a significant difference between the pre-lC and post-lC readings (p=0.0.18). The pre-IC readings were consistently larger than the post-IC readings in both sets of the measurements. For this reason, the remainder of the data analysis was performed on the means of the pre and post V02/kg indirect calorimetiy measurements. There was no significant difference between the two spirometry readings [(p=0.498)(95% Cl = -0.678 to 0.378 mYminlkg)] and therefore the mean of the closed circuit spirometry readings was used in the remainder of the data analysis. The Pearson's correlation coefficient demonstrated a significant correlation between the prelC (r = 0.950,p~O.OOOl) and the post-IC (r = 0.863 p<0.0001) V02/kg measurements when compared to the closed-circuit spirometer V02/kg measurements (Table 2.1). The differences between the two methods, IC and Manuscript written in the fmat of the American Journal of Vetefinary Research spirometry were plotted against their means (Figure 2.1 and 2.2). The level of agreement of 1 ml/min/kg had been chosen as our clinically acceptable level of agreement between these two methods of measuring REE. This level of clinicaf agreement represents approximately 3% of the REE in the healthy dogs. To demonstrate agreement between these methods al1 data points, V02/kg determinations, must fall between the dotted lines & 1 ml/min/kg). There were no data points outside these boundaries, indicating good agreement between these two methods, The mean diîTerence and standard deviation between the two methods was 0.283 + 0.51 mllminlkg for the pre-IC and 0.3 3 + 0.63 mlfminfkg for the post-IC. Reiiabiiity study: The dogs were CO-operativeand appeared relaxed for the majority of readings. The minute expiratory ventilation volumes ranged from 2 to 10 Uminute. However, no attempt was made to record stimuli such as, increased activity or noise in the surrounding area or any muscular movement (tail wagging, lifting of front paw) on the part of the dog that may have increased the minute REE, V02rneasurements and minute ventilation volumes. The lower data points in the individual dogs coincide with sleep or alterations in their state of consciousness. The REE and V02for each series, averaged over individual dogs, on each day is shown in Figure 2.3 and 2.4, The average respiratory quotient (RQ: VC021 V02) for each series on the two days was 0.67. The greatest absolute difference, on average, in REE was 20.1 9 kcallday in series three and the lowest was 1.07 kcallday in series four. Differences of up to 11-8 kcallkglday were apparent on an individual dog basis; however, the majority of these differences were considerably less (Figure 2.5). There was no significant main effect of day or series on the REE (p=0.99, p= 0.82 respectively) or the V02 ( p=0.95 p=0.81 respectively) determinations. However, there were significant two-way interactions (dog'day, dog'series) and a three-way interaction (dog'day'series) in the REE (Addendum:Table 2.4) and V02 (Appendix Manuscript written in the format of the American Journal of Veterinary Research V) detenninations. These interactions indicate that the difference behiveen two dogs or the same dog was not similar between the days studied and the magnitude of this difference was dependent upon which senes the measurement was performed. The differing pattern of RE€ measurements between individual dogs is depicted in Figures 2.5 (Addendum: Figures 2.6 and 2.7). This figure dernonstrates that there was unequal variance from day to day over repeated measures in the same dog. Finally, when al1 other variance components were afike: same dog, same day, same series, there was a relatively large amount of variation which occurred at the minute to minute REE or VOz readings (residual error). The ICC results of REE and V02 are presented in Table 2.2 and 2.3 and are expressed as percentages. There was acceptable reliability of the REE (ICC=0.778) and the V02(ICC=0.786) deteminations over al1 series on the two days. To get a more reliable measure of REE or V02, the first series was discounted and series two and three (REE ICC=0.867 and V02 ICC=0.915 ) or series three and four (RE€ ICC= 0.844 and Von ICC-0.908) were combined. Series three (SD=97.24 kilocalories/day) and four (SD=97.47) have significantly less variance than series one (SD= 149.1 9) and wouid provide a tighter estimate of the mean REE regardless of the day or dog (Figure 2.5). However, in series three and four, when the readings were collected in an alike manner (same day, same dog, same series), 95% of the individual readings can be expected to Vary, above and below the mean, by as much as 195 kilocalories per day or, on average, by 7.27 kcal/kg/day. Dividing the standard deviation by the average weight of the dogs enrolled into the study translated the units to kcal/kg/day. 2.4 Discussion: The agreement study demonstrated that the two indirect calorimetry methods (IC and spirometry) were highly correlated. However, correlation measures association and not agreement; therefore, it cannot assess how interchangeable the two methods may be for assessing an individual patient. The limits of agreement Manuscn'pt wntten in the format of the Arneflcan Journal of Vebnnary Research method of statistical analysis, which was performed in this shidy, is the appropriate test of agreement between two methods of clinical measurement?' The level of agreement was based on previous data in dogs2.25v33v"and a pilot study performed by the primary author on normal dogs. Excellent agreement between the open-flow indirect calonmeter and the closed circuit spirometer was dernonstrated in this study; therefore, either of these two methods can be used interchangeably on an individual basis to measure the VOdkg. However, the ciosed-circuit spirometer is only capable of measuring V02/kg. Furthermore, in order to prevent hypoxemia, an increase in the inspired oxygen mixture is inhaled during this procedure which limits the arnount of collection tirne. Therefore, there may be insufficient time for the body's carbon dioxide and oxygen stores to reach equilibrium when using this method." The open-flow indirect calonmeter, therefore, may be utilised in the clinical setting to obtain a measure of the V02/kg and furthemore an estimate of the REE in the dog. The open-flow indirect calonmeter was able to differentiate a consistent difference between the pre and post IC time periods. This difference may have been due to the presence of secretions and condensation in the tubing which would have resulted in an artifactual decreased in oxygen consumption. If the addition of oxygen into the closed-circuit system, had an effect on the post-IC VOdkg determination an increase in the V02/kg value would have been observed; however, a decreased value was observed in this study. Reliability analysis indicated that despite the inherent variability of the REE readings, the reliability of the serial open-flow indirect calorimetric measurements of REE was acceptable in this study. There was significant variation in individual dogs from series to series and from day to day. However, there was no main effect of day or senes, when averaged over al1 dogs, indicating that there was repeatability of the measures within the group of normal dogs between the two days. The data obtained from the first series behaved differently than the data from the subsequent series. The Manuscript Menin the format of the American Journal of Veterinary Research first series had a diHerent pattern of the variance components and a higher variance- This difference has been observed in other studies involving dogs2 or hum an^'^*'^. The different behaviour of the dogs in Our study may have been due to a leaming effect in the dogs towards the facemask and collection system. Therefore, discounting the first two series and analysing the data in the last two only would obtain the most reliable and precise estimates of REE, A study perfomed by Walters et al2 also demonstrated good reliability of serial indirect calorimetry readings where a total of five evaluations were perfomed in twenty normal dogs over an eight-hour period. The reliability of the energy expenditure per kilogram was better when the middle three readings were averaged ([CC= 0.87), than over ail five evaluations (ICC= 0.02) or when the first evaluation was eliminated (CC= 0.46). The authors concluded that to obtain a reliable estimate of the energy expenditure a 15-minute adaptation period to the collection system should be petforrned and that the first reading should be eliminated using an average of the subsequent readings performed for the data analysis. We were hopeful that the manner in which we collected the reliability data would enable us to measure the RE€ and V02in a steady state. However, no coefficients of variation criteria were provided for the definition of a steady state in the present study. Furthemore, figure 2.5 (Addendum: Figures 2.6 and 2.7) demonstrates that, despite the appearance of a resting and relaxed state in Our dùgs during the procedure, there was still considerable variability in the individual dogs from series to series and from day to day. In human medicine, the definition of a patient's "resting" state is characterised by stable metabolic rneasurements over a predefined time period and is not solely based on the ciinical observation of the patient? A steady metabolic state implies that the gas exchange measurements are equal to tissue gas exchange. It is under these conditions that the assumptions associated with indirect calorimetry are va~id."~' Therefore, some of the variation Manuscript menin Vie format of the Amencan Journal of Veterïnary Research obsewed, and the lower RQ's, rnay have been due in part to insufficient measurement time required to obtain a steady state in Our dogs despite their apparent state of rest, Also this open-fiow indirect caiorimetry system has previously shown a tendency to underestirnate the VC02 at lower expired ventilation volumes in humans? Therefore, at low ventilation volumes. as were observed in this study (2 to 10 Umin), an underestimation of the VC02 would be more likely to occur resulting in the lower RQ values. This may have been due to the proportional sampling technique that was utilised by the indirect calonmeter as previously described.*' However, even if large errors in the distribution of energy production among the three fuels (fats, carbohydrates and protein) were present, there would have been litüe effect on the total energy expendikire because the caloric equivalent of oxygen varies by less than 15% The range of V02 values in this study (4.1 5 to 6.40 mi/min/kg) were comparable to those reported by others in resting dogs (6.0 to 12.3 rnl~min/kg).~*~~*~~*~~The range of the energy expenditure values was also comparable (17.28 to 63.62 kcal/kg/day) to those reported by Walters et al (38.72 to 50.42 kca~/kglday)~and Galvao (31 -44 to 57.84 kcallkglday)? Some of the variability in the ranges observed in our study may be due to the level of consciousness in the dog during the test procedures. Peters et ala has also reported this level of fluctuation in V02values measured by indirect calorirnetry in dogs. In their study there was a substantial difference in the means of the V02values, which were dependent upon the level of awareness of the dogs. In a resting state the V02values fiuctuated between 2 to 10 rnllmirdkg with a rnean of 5.57 + 1.18 (SD). The rnean V02value decreased to 3.97 + 1-01 ml/rnin/kg in a drowsy state and was at its lowest value and least variable when the dogs were asleep (2.46 + 0.48 ml/min/kg). Therefore, it appears important that when the measurement of oxygen consumption is perfomed Manuscript menin the format of the American Journal of Veterinary Research the state of awareness is recorded and atternpts made to ensure that the subjects are in a similar state of awareness throughout the procedure. In conclusion, we ernphasise that the two indirect calorimetry methods of measuring VOdkg in this study, via open-flow indirect calorimetry or closed-circuit spirometry, can be said to agree. Therefore, the open-flow indirect calorimeter rnay be utilised in the clinical setting to obtain a measure of the V02 and, moreover, an estirnate of the RE€ in the dog. The measurement of the REE and V0& also reliable within a group of normal dogs between two days. This is particularly tnre if, after a 10- minute period of adaptation, the first series is discounted and multiple series are perfonned with the last two serial measures being utilised for the REE detemination. Manuscript written in the fonnat of the Arnerican Journal of Veterinary Research Footnotes: a TEEM 100 Metabolic apparatusa, Aerosport inc, Ann Arbor, USA. Benedict-Roth spirorneter, Warren E. Collins Inc, Braintree, USA. " Atravet, Ayrest laboratories, Montreal Canada. Torbugesic, Ayrest laboratories, Montreal Canada. Rapinovet, Schering-Plough, Pointe-Claire, Canada, ' Aquamatic, Amencan medical Systerns, Valencia .USA Eukanuba Maintenance, larns Company, Dayton, USA. Ham Rudloph Inc., Kanas City,USA. 'SAS Institute, Cary, NC. Manuscript written in the format of Vie Amencan Journal of Veterinary Research Figure 2.1 .The difference between the two indirect côlorimetry methods; pre-open- fiow indirect calorimetry (IC) and the closed-circuit spirometer (SP), for the determination of oxygen consurnption per kilograrn 0(02/kg) are plotted against their means. To demonstrate that pre-IC and SP agree ail data points should lie between the dotted lines. ., Difference between pre-IC VOa/kg and SP VOa/kg. Manuscript written in the format of the Amencan Journal of Vetennary Research Figure 2.2. The difference between the two indirect caforimetry methods; post open- flow indirect calorimetry (IC) and the closed-circuit spirometer (SP), for the determination of oxygen consumption per kilogram NO2 /kg) are plotted against their means. To demonstrate that pre-IC and SP agree al1 data points should lie between the dotted lines. m. Difference between post-IC V02/kg and SP V02 /kg. Manuscript written in the format of the Arnerican Journal of Veterinary Research 2 3 4 Se ries Figure 2.3. Mean + SEM resting energy expenditure (kcal/day) via the open-flow indirect calorimeter in staff owned dogs (n=7) evaluated over four 16-minute series over the two days. A series is comprised of a IO-minute adaptation period to the face mask and collection system and six one-minute readings for the data collection. Manuscript Menin the format of the American Journal ofveterinary Researcfi +Day 1 -I-- Day 2 Figure 2.4. Mean 2 SEM oxygen consumption (WOZmllmin) via the open-fiow indirect calonmeter in staff owned dogs (n=7) evaluated over four 1&minute series over the two days. A series is comprised of a 10-minute adaptation period to the face mask and collection system and six one-minute readings for the data collection. Manuscript written in the format of the American Journal of Veterinary Research &Dog 1 +Dog 2 - -a- - DO^ 3 4- Dog 4 --a-- DO^ 5 &Dog 6 --A- DO^ 7 Se ries Figure 2.5. Difference in REE determii:ed by open-flow indirect calorirnetry, Lietween the days, over the four series in individual dogs. The data points for each dog represent the best linear unbiased estimate (BLUPS) of the REE. The BLUPS are the best estimate of what the actual RE€ will be for an individual when al1 the different variance components had been taken into account. Manuscnpi written in the format of the Amencan Journal of Veterinary Research Table 2.1 .Pearson correlation coefficients and probability values calculated for the pre and post open-flow indirect calorimetry oxygen consumption per kilogram (V02/KG) readings compared to the closed-circuit spirornetry VO$KG reading. Normal dogs (n=7) Correlation Coefficient Probability Value Pre-l ndirect calorirnetry 0.95 p<0.0001 Post-Indirect calorirnetry 0.86 p Manuscript written in the format of the Amencan Journal of Veterinary Research Table 2.2: lndicates the interclass correlation coefficient for the resting energy expenditure (REE) of the averaged six minute readings expressed as a percentage (Le. ICC value x 100). The rernainder of the table indicates the percentage of the total variance which is attributed to each of the different lev& for example: the dog, the series, the residual error and one of the significant interactions (dog* day) among these variables for REE measures. LeveI of variance for the Series 1-4 Series 24 Series 2-3 Series 34 REE % % % % DOG 22.8 41 -8 49.1 46.6 DOG'DAY 27.0 24.7 17 20.7 SERIES interactions 35.7 21.6 22.7 21-9 RESIDUAL ERROR 22.6 15.1 14.4 13.8 Average of 6-M I Na 77.9 90.4 86.7 91-1 Residuai Error Variance 12513.03 9346.46 9291.8 9398.43 Average of 6-MIN 9459.1 4 5017.0 7328.77 5247.74 Error Variance a 80% and above = good reliability.70-79%= acceptable reliabilty.69% and below = poor reliability These values are the actual residual error variance components as estimated by the proc mixed model. Manuscript menin tfie fmat of the American Journal of Veterinary Reseorch Tabie 2.3: lndicates the interclass correlation coefficient for oxygen consumption NO2)of the averaged six minute readings expressed as a percentage ( Le. ICC value x 100)- The remainder of the table indicates the percentage of the total variance which is attnbuted to each of the different levels, for example: the dog, the series, the residual error and one of the significant interactions (dog* day) among these variables for the V02 measures. Level of variance for Series 14 Series 2-4 Series 2-3 Series 34 Oxygen Consum ption % % % % DOG 23.7 43.7 50.9 48.5 SER1 ES interactions 36.4 21.l 21.5 21.2 RESIDUAL ERROR 25.6 16.2 15.8 14 Average of 6-MIN a 78.6 89.4 86.9 89.1 Residual Error 309.95 217.40 218.27 205.59 Variance Average of 6-MIN 192.65 119.38 152.31 137.82 Error Variance a 80% and above = good reliabiIity.70-79%= acceptable reliabilty.69% and below = poor reliability These values are the actual residual error variance components as estimated by the proc mixed model. Manuscript menin the fonnat of the American Journal of Veten'nary Research 71 2.5 Reference List: McClave SA, Snider HL. Use of indirect calorimetry in clinical nutrition. NCP 1992; 7:207-221. Walters LM, Ogilvie GK, SaIman MD, Fettman MJ , Joy J, Hand MS, Wheeler SL. Repeatability of energy expenditure rneasurement in clinically normal dogs by use of indirect calon'rnetry. Am J Vet Res 1993; 54:188l -q 885. Weir JB. New methods for calculating metabolic rate with special reference to protein metabolism. J Appl Physiol 1949; 109:l-9. Mann S. Westenskow DR, Houtchens BA. Measured and predicted caloric expenditure in the acutely ill. Crit Care Med 1985; l3:l73-l77. Weissrnan C, Kemper M, Askanazi J, Hyman Al, Kinney JM. Resting rnetabolic rate of the cntically il1 patient: Measured verses predicted. Anaesthesiology 1986; 641673-679. Kinney JM, Duke JH, Long CL, et al. Tissue fuel and weight loss after injury. J Clin Path 1970; 23:65-72. Elwyn DH. Nutritional requirements of adult surgical patients. Crit Care Med 1980; 819-20. Blaxter KL. The minimal metabolism. 1n:Blaxter KL ,eds. Energy mefabolism in animals and man. Cam bridge: Cam bridge University Press, 1989; 120-146. Feurer 1, Mullen JL. Bedside measurernent of resting energy expenditure and respiratory quotient via indirect calorirnetry. NCP 1986; 1 :4346. 10. Bursztein S, Elwyn DH, Askanazi J, Kinney JM , Kvetan VI Rothkopf MM, Weissman C. lndirect caIorimetry: History and overview. 1n:Bursztein S, Elwyn DH, As kanazi J ,et al ,eds. Energy Metabolism, lndirect Calonmetry, and Nutrition. Baltimore, Maryland: Willarns & Wilkins, 1989; 2-23. 11. McClave SA. Spain DA. Indirect calorimetry should be used. NCP 1998; 13:143- 145. 12. Zavala DC. Evaluating nutritional needseds. Nutritional Assessment in Crifical care: A Training Handbook. lowa City: University of lowa press, 1989; 41-59. Manuscript written in the format of the American Journal of Veteflnary Research f3. Isbell TR, Klesges RC, Meyers AW, Klesges LM . Measurements reliability and reactivity using repeated measurements of resting energy expenditures with a facemask, mouth piece and ventilated canopy. J Parenter Entera1 Nufr 1991; l5:165-168. 14. Leff ML, Hill JO, Yates AA, Cotsonis GA, Heymsfield SB. Resting metabolic rate: Measurement reliability. J Parenter Entera1 Nutr 1987; 1 1:354-359. 15. Verrneiji CG, Feenstra BWA, Vanlanschot JJB, Bruining HA. Day-to-day variability of energy expenditure in critically il1 surgical patients, Crit Care Med 1989; 17:623-626. 16. Smyrnios NA, Curley FJ, Shaker KG. Accuracy of 3û-minute indirect calon'metry studies in predicting 24-hour energy expenditure in mechanically ventilated, critically il1 patients. J Parenter Entera1 Nutr 1 997; 21 :l68-174. 17. Head CA, McManus CB, Seitz S, Grossman GD, Stanton GW, Heymsfield SB. A simple and accurate indirect calorimetry system for assessment of resting energy expenditure. J Parenfer Enteral Nutr 1984; 8:45-48. 18. Brandi LS, Grana M, et al. Energy expenditure and gas exchange measurements in the postoperative patients: thermodilution versus indirect calorimetry. Crif Care Med 1992; 20:1273. 19. Ogawa AM, Shikora SA, Burke LM, Heetderks-Cox JE, Bergen CT, Muskat PC. The thermodilution technique for measuring resting energy expenditure does not agree with indirect calorimetry for the critidly il1 patient. J Parenter Enteral Nufr 2000; 22~347-351. 20. Bursztein S, Saphar P, Singer P, Elwyn DH. A mathematical analysis of indirect calorimefry measurernents in acutely il1 patients. Am J Clin Nutr 1984; 50:227- 230. 21. McClave SA, Snider HL, Greene L, et al. Effective utlization of indirect calorimetry during critical care. Intensive Care Worid 1992; 9: 194-200. 22. Brandi LS, Bertolini R, Calafa M. Indirect calorimetry in critically il1 patients: Clinical applications and practical advice . Nutrition 1997; 13549-358. 23. Feurer 1, Crosby LO, Mullen JL. Measured and predicted resting energy expenditure in clinically stable patients. Clin Nufr 1984; 3:27-34. Manuscript menin the fonnat of the American Journal of Veterinary Research 24. Walton RS, Wingfield WEI Ogilvie GK, Fettman MJ, Matteson VL. Energy expenditure in 104 postoperative and traumatically injured dogs with indirect calorimetry. J Vet Emerg Crif Care 1998; 6:71-79. 25. Ogilvie GK, Salman MD, Kesel LI Fettman MJ. Effect of anaesthesia and surgery on energy expenditure determined by indirect calorimetry in dogs with malignant and non-malignant conditions. Am J Vet Res 1996; 57:1321-1325. 26. Bursztein S, Elwyn DH, Askanazi J, Kinney JM . Theoretical framework of indirect calorimetry and energy balance.eds. Energy Metabolism, Indirecf Calonmetry and Nurifion. Baltimore, Maryland: Williams & Wilkens, 1989; 27-81. 27. Canadian Council on Animal Care. Guide to the care and use of experimental animais volumes 1 and 2, in 1984. 28. Operatot's Manual, Aerosport Inc. TEEM 100 Metabolic analysis system total energy expenditure measurement. TEEM 700 Metabolic Analysis System Operator's Manual 1993; 1-47. 29. Westenskow DR, Schipke C.A., Raymond JL, Saffle JR, Becker JM, et al. Calculations of rnetabolic expenditure and substrate utilization from gas exchange measurements. J Parenter Enteral Nutr 1988; l2:20-24. 30. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods. Lancet 1986; i:307. 31 . Winer WB. Statistical principles in experimental design. New York :McGraw-Hill , 1971. 32. Searle SR, Casella G, McCulloch CE. Variance components, New York :John Wiley and Sons Inc, 1992. 33. Galvao PE. Heat production in relation to body weight and body surface.lnapplicability of the surface law on dogs of the tropical zone. Am J Physiol 1946; 148:478490. 34. Peters J, Arndt AO. Letter to the editor: Fluctuations in resting energy expenditure. Crif Care Med 1984; 926. 35. Weissman C. Oxygen transporfation and ufilisation. Fulierton Ca :The Society of Critical Care Medicine, 1987;25-64. Manuscript written in the format of the Arnerican Journal of Vetennary Research 36. Weissman C, Kemper M, Damask MC, et al. Effect of routine intensive care interactions on rnetabolic rate. Chest 1984; 86:815-818. 37, Ferrannini E. The theoretical basis of indirect calorimetry: A review. Metabolis 1988; 37~287-301. 38. Wideman L, Sioudemire NM, Pass KA, McGinnes CL, Gaesser GA, WeItman . Assessment of the Aerosport TEEM 100 portable rnetabolic measurement system. Med Sci Sport Exer 1996; 28:509-515. 39. Francisco R, Aquin L, Searles JM, Banchero N. Oxygen cost of breathing in dogs. Respiration 1978; 35: 186-1 91. Manuscript menin the fmat of the Amencan Journal ofvetennary Research 2.6 Addendum for Manuscn'pt 1: 1400 - +Dog 1 1300 - +Dog 2 1200 - -O- M DO^ 3 4-Dog 4 1100 - --Q--Dog5 W. 0" &Dog 6 c 1000 - O A -A- Dog 7 2-b S 900 - 5 Y 800 - W W CE . . 700 - ,' .. r 600 - B - * 500 - 400 % 8 1 O 1 2 3 4 5 Series Figure 2.6. The REE (kcdday) detemined by open-flow indirect calorimetry, of each dog over the four series on day 1. The data points for each dog represent the best Iinear unbiased estimate (BLUPS) of the RE€. The BLUPS are the best estimate of what the actual REE will be for an indiviaual when al1 the different variance components had been taken into account. Manuscn'pt written in the fonnat of the Arnencan Journal of Veterinary Research Figure 2.7. The REE (kcalfday) deterrnined by open-flow indirect calorimetry, of each dog over the four series on day 2. The data points for each dog represent the best linear unbiased estirnate (BLUPS) of the REE. The BLUPS are the best estimate of what the actual REE will be for an individual when al1 the different variance components had been taken into account. Manuscript written in the format of the Amencan Journal of Veterinary Research Table 2.4. Estimate of the variance components associated with the resting energy expenditure (REE) measurements over the hnro days. The data in the estimate colurnn was used to compute the interclass correlation coefficients for the REE over the four series. Level I Estimate Std Error Probability Value DOG DOG*DAY DOG*SERIES DAY*SERIES DOG'DAY'SERIES Residual Manuscriptwrittenin the format of the American Journal of Veterinary Research 78 Cha~ter3: Cornparison of Indirect Calorimetry and the Predictive Equation for the determination of Resting Energy Expenditure in il1 and healthy dogs. 3.0 Abstract: Objective: The objective of this study was to detemine the level of clinical agreement for resting energy expenditure (REE) measured by open-flow indirect calorimetry (MREE) and the REE predicted by the commonly used predictive equation (PREE) in healthy and il1 dogs. Design: Prospective case series. Animals: 77 dogs were enrolled: Two groups were evaluated. The first group consisted of 1O healthy staff-owned control dogs. The second group consisted of 67 client-owned dogs. Thirty-four with significant medical illnesses, 16 recovering from major surgery and 17 experienced a traumatic incident 24-72 hours prior to presentation. Procedure: The oxygen consumption (V02) and carbon dioxide production ( VC02 ) in Umin was measured at three-time periods (1 1:O04 4:OO, 16:OO-20:00,03:00-06:OO) for 15 minutes during a 24-hr period, for two days, using open-flow indirect calorimetry and a facemask. The abbreviated Weir formula r(3.94 (VOz) +1.l (VC02) x 14401 was used to calculate the 24-hour EE from the calorimetric data. A consecutive five minute period in the 15-minute reading, where the coefficient of variation was 10% or less in the il1 dogs or 15% or less in the healthy dogs and the RQ was between 0.67 to 1.3, was used to determine the V02and VC02values. Results: Mean values of the MREE versus the PREE in healthy dogs and those with medical diseases were not significantiy different. However, there was a trend towards overestimation of the MREE by the PREE in the group of traumatised dogs on day 1 (p=0.06) and there was a significant difference between the MREE and PREE in the postoperative group on day 2 (p=0.018). The PREE overestimated the MREE, on average, by 20.95% in this group. Moreover, there was great variation between these two methods on an individual dog basis such that, on day 1 and day 2 respectively Manuscript menin the format of the Amencan JoumaI of Veterinary Medicine 79 49% and 43% of the dogs' individual REE determinations differed by 2 40%. Furthemore, the PREE will only agree to within 5 20% of the MREE between 510Io to 57% of the time in individuals of any group on either day. Conclusions: The level of agreement between these two rnethods for determining the 24-hour REE was clinically unacceptable. On average, the predicted equation can predict the rneasured REE in the healthy dogs and those with medical illness. However, the equation, on average, will result in values 15 to 20% above the MREE in the dogs with trauma or major surgery. Furthemore, in an individual, the predictive equation will only be accurate to within + 20% of the measured energy expenditure approxirnately 45% of the time. In other words approxirnately 45% of the critically il1 dogs' energy needs cannot accurately be estimated if the traditional predictive equation for energy expenditure is utilised. Clinical Relevance: The level of disagreement between the two methods studied, the measured REE and the predicted REE, indicates that the importance of accurately estimating the REE in il1 or healthy dogs by the predictive equation is unacceptable. 3.1 Introduction: Advancements in veterinary medicine have enabled clinicians to successkilly treat more challenging and complex medical and surgical problems. Nutritional support plays an integral role in recovery from these illnesses or injuries. If adequate caloric intake is not provided during the treatment and recovery phases of illness, the risk of morbidity and mortality increases.'" During illness and injury, energy and substrates are mobilised from lean body mass in an attempt to support inflammation, immune function and tissue repair. '*34 The degree of mobilisation of endogenous nutrients and the increase in the energy expenditure (EE kcallday) depends on the type and degree of illness or injury? The severity of the catabolic state is such that between 3040% of hospitalised hurnan patients are in a state of protein-calorie ma~nutrition.~~.Adequate nutritional support may help to reduce the catabolic process and maintain the body's Manuscript written in the format of the American Journal of Veterinary Medicine 80 lean mass dunng this period of altered rnetabo~isrn.'~~*~However, excessive nutrient intake may place additional stress on the respiratory system9, cardiovascular system1° and liver." Both under and overfeeding critically il1 human patients can therefore impact negatiiely on clinical d ut corne.'^^'^ It has been assumed that dogs respond in a simiiar manner as humans do to illness and disease; severely il1 or injured dogs have an increased resting energy expenditure (REE kcal/day) which is proportional to the severity of the injury. Catabolisrn of endogenous nutrient stores is increased to meet the energy demands associated with illness, which in turn predisposes the animal to protein- calorie malnutrition. These assumptions are prevalent in the veterinary literature13; 14however, validation is sparse. 15-16 In veterinary patients energy requirements are calculated using a predictive equation which atternpts to estimate the daily energy expenditure. The predictive equation is expressed as PREE = 70 x (BW in and is multiplied by a stress or illness factor. The illness factor is a subjective correction factor which ranges from 1.25 to 2 and is meant to take into account the increased REE associated with injury or illness. 13-14The illness factors have been derived rnainly from clinical data in humans l7rather than from clinical or experirnental data in dogs. l5 Kleiber and Brody derived the equation: 70 x (BW in , in 1932 by studying mature animals of 26 different species in a respiratory calonmeter at "rest", or at least without abnormal activity, in the postabsorptive state. " The equation 70 x (BW in kg)'-.'' is the best fitting line of regression of a log-log plot of the resting metaboiic rate of species ranging in size from a mouse to an elephant. The predictive equation of the daily EE is based on the assumption that the body weight in kilograrns to a power function of the 0.75 can accurately reflect the size, and therefore the metabolic activity, of the body cell mass. It is unclear from the veterinary literature when or why this equation was adopted and as of yet, it has not been verified in il1 or Manuscript written in the format of the Arnerican Journal of Veterinary Medicine 81 injured dogs. In fact, there is compelling evidence to suggest that this equation is not accurate in healthy dog~.'~*~ The predictive equations for the estimation of REE in human medicine, even when corrected for stress and illness factors, differ significantfy frorn the measured REE.~'-'~Therefore, a more reliable approach to assessing the daily EE is to measure the EE under a standardised set of conditions. As it is impractical to achieve basal conditions for the measurement of EE in hospitalised patients (defined as; first thing in the moming, after an ovemight sleep, fasted for at Ieast 10 hours, with a normal body temperature, in a supine position and in a thermoneutral environment with no physiologic or psychological stress), energy expenditure is measured under resting conditions. The "resting" conditions under which REE is measured are less strictly defi~~ed~~and may include the following situations: diet-induced thermogenesis. physiologic or psychological stress factors and variations in body and environmental temperature. It is important that a clear definition of the conditions under which the metabolic rate is measured be provided in the experimental protocol so that compatisons arnongst researchers may be performed? Indirect calorirnetry, which measures gas exchange [oxygen consumption 0(02)and carbon dioxide production (VC02)] at the level of the lungs, is currently the method of choice in measuring REE in hospitalised These gas exchange measurements are converted to kilocalories per day by using the abbreviated Weir formula or by calculating the calonc equivalent of the litres of oxygen consurned and carbon dioxide produced with the use of standardised tables."*3042 The use of indirect calorimetry for assessing REE in dogs has only recently been examined in veterinary medi~ine.'~@~~~The importance of accurately and reliably assessing the daily EE of ili or injured animals is pivotal to maxirnising the beneficial effects of nutritional support. Examination of the level of agreement between the veterinary standard of care for estimating the daily EE and the measurement of REE by indirect calorimetry would have a notable clinical impact in veterinary medicine. Manuscript written in the format of the American Journal of Veterinary Medicine 82 The objective of this study was to assess the level of clinical agreement between open-fiow indirect calorimetry and the traditional predictive equation for the 24-hour energy expenditure (EE) in a population of healthy dogs and of il1 dogs in an ICU setting at a Veterinary Teaching Hospital. These conditions of hospitalisation would be similar to those reported in human indirect calorimetry st~dies.*~~~~**~~We hypothesised that the REE as measured by indirect calorimetry woufd not agree with the RE€ estimated by the predictive equation. 3.2 Material and Methods: The study population consisted of 77 dogs evaluated at the Ontario Veterinary College Teaching Hospital between April1999 to September 1999. The 77 dogs included, 1O staff-owned healthy dog which constituted the control population, 34 dogs with medical illnesses, 17 dogs recovenng from major surgery and 16 with major trauma. The traumatic incident occurred 24-72 hours pnor to presentation. Clinical Cases: These dogs were enrolled based on presentation order and were included in the study if: co-operative, within 10% of ideal body weight and would stay in the intensive care unit for a minimum of 48 hours (72 hours if they required a general anaesthetic). Exclusion criteria were: uncooperative patient, greater than 10% of ideal body weight, requirement for supplemental oxygen, facial conformation that would not allow for an-air-tight seal with the face rnask, discharged from the ICU or euthanised prior to completion of at least three readings of REE by the indirect dorimeter, or a suspected or confirrned diagnosis of neoplasia. The 67 dogs were divided into three groups according to illness or injury. Dogs in the medicine, post- operative or trauma groupings were those dogs who required extensive monitoring and care in the ICU, for a minimum of 48 hours or 72 hours if they undenvent a general anaesthetic. Dogs requiring a surgical procedure secondary to a traumatic incident were considered as a member of the trauma group and were not corn bined into the postoperative group. The clinician responsible for each case made decisions Manuscn'pt writîen in the format of the Amencan Journal ofveterinary Medicine 83 regarding therapy and caloric intake entirely independent of the study; however, al1 il1 dogs were fed a similar diep. Confrol Group: The control population of dogs was obtained frorn the staff population at the Ontario Veterinary College. Dogs were judged to be clinically healthy based upon physical exarnination, hernograrn, biochernical profile, and free serum thyroxine concentrations being within normal limits. In addition, there was no previous rnedical history or known exposure to anaesthesia or exogenous corticosteriods within the previous rnonth. All dogs were within 10% of ideal body weight and were fed a high quality commercial dog foodb on a sirnilar feeding schedule. The control dogs were included in this study in order to provide a comparative group for the il1 dogs in this study and for those previously published REE values in the dog and to assess the level of vatiability present in healthy dogs MREE. Protocols were followed in accordance with the guidelines set by the Canadian Council on Animal care4' and were approved by the cornmittee concemed with ethical review at the University of Guelph. lnformed consent was obtained pnor to enrolrnent into the study. Determination of the measured resfing energy expendifure (MREE): Upon entry into the intensive care unit and enrolrnent into the study blood was collected for: a complete blood count, serum biochemical profile and bIood gas analysis. The dogs were provided with a minimum six-hour period to acclimatise to the intensive care unit surroundings pnor to the first calonmeter reading being performed. If general anaesthetic was required in any of the dogs, a hl1 24-hour recuperation period was provided prior to performing rnetabolic rate measurements. The measurernent petiods during the 24-hours were perfomed at the following tirnes: O3:OO-O6:OO,ll:OO-l4:OO, l6:OO-2O:OO on two days. Al1 metabolic measurernents were performed at the cageside in the intensive care unit, at least 30- minutes after administration of analgesic rnedication if required, 30 to 60 minutes after a diagnostic or therapeutic procedure or physical activity (walks) and three to four hours after a Manuscript written in the format of the American Journal of Veterinary Medicine 84 meal. Dogs enrolled in the study were fed a similar dietd twice a day (08:OO and 20:OO). If the dogs were receiving partial parenteral nutrition the rate of administration was not altered in the 12 hours prior to metabolic measurements. Control Group: The control dogs were hospitalised and cage rested for their measurements, which were perfotmed during similar time periods as the il1 dogs. However, they were performed in a quiet thermoneutral room within the hospital and not within the ICU and the diet fed consisted of a commercial maintenance foodb. Only five of the healthy dogs in the control group had the indirect calorimetry readings perfonned on Wo days. A 1O-minute adaptation period to the facemask and collection system was performed prior to each measurement. A reading consisted of 15 consecutive one- minute V02and VC02 measurements. A steady state was defined as a consecutive five minute reading in which the coefficient of variation of the V02 and VC02 was 5 10% in the il1 dogs and 5 15% in the healthy dogs and where the respiratory quotient (RQ) was within the physiologie range (0.67-1.3). Defermination of fhe predicted resfing energy expenditure (PREE): The equation. 70 x (BW in kg)O-" x illness factor was utilised to determine the predicted REE(PREE). The weight of the dog in kilograms at entry into the study was utilised for the calculation of the PREE. The calculation was perfomed at the time of enrolment. An illness factor was determined for each dog based on the initiaf clinical signs and physical examination. The patient's illness factor was assigned based on the following criteria: cage rest = 1.25, post-surgery = 1.25 to 1.35, trauma = 1.35 to 1.5, sepsis = 1.5 to 1.7 (Appendix VIII).The illness factor was determined by the primary author (eot) at the time of enrolment prior to any metabolic measurernents being performed. Control Group: The illness facto for al1 control dogs was considered to be 1.O. Measured and predicted REE were expressed in kilocalories/kg/day. Manuscript written in the format of the Amencan Journal afveterinary Medicine 85 Indhct calorimefry= Measurements of V02 and VC02via the portable open-fiow indirect calonmete? were performed by the primary author (eot), a trained research assistant or a trained ICU technician. The first two individuals performed the rnajority of the readings (48.5 % and 24.2 % respectively). The instrument measures the partial pressure of oxygen and carbon dioxide in expired air by a galvanic fuel cell and a nondispersive infrared analyser, respectively. Expiratory ventilation volume is measured with a flat plate orifice pneumotach. The flow of gas through the sharp edged, thin plate pneumotach onfice produces a pressure drop proportional to the square root of the flow. This flow signal is then integrated to obtain the expired air volume. lndependent instruments within the system measure temperature, barornetric pressure and time. The rnicroprocessor cafculates, using standard equations, V02, VC02 and respiratory quotient (RQ). " Before use. the gas analysen were calibrated to a standard gas of known composition (18.0 % O2and 2.0% COp)on a daily basis. The pneumotach was calibrated with a 1-L calibration syringeCon a weekly basis. Measurements of V02and VC02 in Umin were calculated every 20 seconds and an average of these 20-second readings were pn'nted at one-minute intervals. The REE (kcdday) was calculated by using the abbreviated Weir formula: 13-94 0/02) + 1-1 (VC02)]X 1440, where VOz= Onconsumption (Umin) and VC02 = CO2production (Umin). Sfatistical Analysis: The data set analysed only included those indirect calorimetry readings in which the RQ in the five-minute steady state reading were within the physiologic range (0.67-1 -3). The MREE readings per day were averaged to give a mean REE measurernent in kcal/day, which was then compared to the PREE in kcal/day. Values obtained from the measured REE were assessed for normal distribution using a Wilk- Shapiro test. The extent to which the MREE and .PREE differed per day was calculated as a logarithm of the ratio MREEI PREE (i.e.log MREE/PREE=log MREE- Manuscript written in the fwmat of the Amencan Journal af Vetennary Medicine 86 log PREE). A power calculation (1-p =0.8) for the sample size per group was performed so that if a 30% difference per day was present between the MREE and PREE it could be detected (Le. logarithm of the ratio=1.34 or 0.7408). Statistical analysis of the ratio was performed using a general linear mixed model, which âssessed the random effects of the dogs, categorical variables of the group, day, gender, breed and continuous variables which were anaIysed by regression analysis. The data set was analysed using proc mixede4', which tested whether the means of the groups' logamhm ratios were different from one. In addition an F-test was used to detect which of the covariates; group, day, age, breed, gender and any interaction amongst or between these variables, had sig~ificantinfluences on these means. Other covariants such as, body temperature, heart rate, respiratory rate, serum carbon dioxide and serurn bicarbonate concentration were also assessed in those individuais in which this information was available( n= 97 readings). Variables were retained in the model if they were significant at 10% (p=0.1). The primary endpoint of the statistical analysis is to assess whether there is a significant ditference between the MREE and PREE via the ratio of these two variables (i.e. if MREE and PREE are different the ratio is significantly different frorn one). The logarithm of the ratio gives a relative percentage change between the two methods and not the absolute difference. Therefore, if the logarithm of the ratio is positive the MREE is a relative percentage greater than the PREE and if the logarithm of the ratio is negative then the MREE is a relative percentage less than the PREE. This statisticaf analysis examines the means of the logarithm ratio for each group on two different days and does not examine how well these two methods would agree for an individual dog. A statistical analysis of the level of clinical agreement between the predicted and measured REE for each dog is also perfor~ned.~~A natural logarithmic transformation of the data was performed. The difference of the natural log of the REE determinations was graphed against the natural log of the mean of the REE determinations. Therefore, an antilog function must be performed so that the values on Manusdpt written in the format of the American Journal ofveterinary Medicine 87 the graphs reported may be reverted back into the original units of kcallday. A level of clinical agreement of + 20% difference per day was deemed clinically significant. In other words if the two methods for the determination of REE; MREE and PREE, differed by more than 20% per day in an individual dog then there would be serious clinical implications by choosing one method over the other. A commercial statistical software program" was used to perform the data analysis. The adjusted or least square means and the SEM are reported. The adjusted means are means which take into account the covariates; group, day, age, group*day. The influence of outlying data points was assessed by removal of these data points and re-running the model. A 0.05 level of probability was used. 3.3 Results: A total of 384 readings were performed in the 77 dogs; however, only 231 readings met our inclusion criteria. The unadjusted mean 5 SD of the age, weight, illness factor and RQ for each group is displayed in Table 3.1 and the gender distribution in Table 3.2. A variety of breeds were enrolled into this study with mixed breeds at 19.9% and Labrador Retrievers at 10.2% most highly represented (Appendix 6.7). Fifteen of the dogs in the medicine group had an immune-rnediated disease (immune-mediated thrombocytopenia, immune-rnediated haernolytic anemia, polyarthritis), nine an infectious disease, three intoxication, three pancreatitis, two renal disease and the remainder had a fever of unknown origin or encephalitis. Thirteen of the dogs in the post-operative group had an exploratory laparotomy, two dogs a neurosurgical procedure, one dog a thoracotomy and one dog an orthopaedic procedure. In the trauma group injuries were caused by a motor vehicle accident in fourteen of the dogs and large dog-srnall dog altercations in two dogs (Appendices 6.7 and 6.8). The logarithmic transformation of the data used to obtain the relative percentage difference per day between the two methods, MREE and PREE, also normalised the data such that the ANOVA assumptions appeared to be adequately met. The adjusted mean and 95% confidence interval of the MREE and PREE in Manuscript written in the format of the Amencan Journal of Vetefinary Medicine 88 kcal/kg/day are displayed in figure 3.1 (Addendum:Table 3.4 and 3.5). The adjusted mean takes into acwunt the effects of the covariants in the model (group, age, day, group'day). A sample size of 30 dogs per group was required in order to detect a significant difference between the MREE and PREE of 30% at the 0.05 level. A sample size of 80 dogs per group would have been required to demonstrate a statistically significant 20% difference per day between the MREE and PREE at the 0.05 level. There was no main effect of group (p=0.186), sex, breed or any interaction amongst or between these variables on the logarithm of the ratios means. However, there was a foced effect of age (p=0.043), day (p=0.0345) and a random effect of the individual performing the test. The MREE declines by 2.58% per year of age and accounts for 10.87% of the total variation observed. The variabIelndaynaccounts for 14.6% of the total variation obse~edand 1% of the overall variation in the MREE could be attrÏbuted to the individual performing the reading. The adjusted mean value of the MREE versus the adjusted mean PREE in healthy dogs or dogs with medical diseases or in the post- traumatic period were not significantly different on either day. The ratio (MREUPREE) was not significantly diHerent from one at any time. There was a trend towards significance in the trauma group's ratio on day 1 (p= 0.06). The trauma group's adjusted mean MREE tended to be on average 15.5% less than the adjusted mean PREE on day 1. There was a significant difference per day between the MREE and PREE (Le. the logarithm of the ratio was significantly different from one) in dogs in the post-operative group on the second day (p=0.0181). The adjusted mean MREE on average was 20.95% less than the adjusted mean PREE in the postoperative group on day 2 (Figure 3.1). In figure 3.2 and 3.3 the differences between the MREE and the PREE in each dog is plotted against their means. To show agreement between the MREE and PREE in an individual dog, al1 data points (REEdeterrninations) must faIl between the clinicaIly acceptable level of agreement (dotted lines) at + 20%. Many data points Manuscript Menin the format of the American Journal d Veterinary Medicine 89 (Figures 3.2 and 3.3, Table 3.3 and Appendix 6.9) lie outside these Iirnits, which demonstrates poor agreement between the MREE and PREE on an individual dog basis in each-ofthe four groups on either day- Moreover, on day 1 and day 2 respectively, 49.2% and 43.2 %, of the dogs' individual REE detenninations by the two methods differed by 140% per day (Table 3.3). There is wide variation in the REE determinations, such that on day 1, the 95% confidence interval (Figure 3.2; the Mean -+ 2SD lines) of the mean difference in the REE determinations as assessed by the PREE would be lower than the MREE by as much as 82.4% or greater than the MREE by as much as 50.4%. On day 2, the 95% confidence interval (Figure 3.3; the Mean + 2SD lines) of the mean difference of the REE determinations as assessed by the PREE would be lower than the MREE by as much as 82% or greater than the MREE by as much as 45.8 %. However, there was no consistent bias observed when the two methods; MREE and PREE were compared. The PREE tended to be greater than the MREE, on average, when al1 groups were combined, by 0.6% on day 1 to 4.88% on day 2 (Figures 3.2 and 3.3). However, when the groups were examined individually on each day a consistent bias was not observed. For example the PREE was lower than the MREE, on average, on day 1 in the control group by 5.75%. whereas, on day 2 the PREE was greater than the MREE by 15.2%. The PREE tended, on average, to be greater than the MREE in the traumatised and surgery dogs. However, the relative percent difference per day was not consistent over the days in each group. For example the PREE was greater than the MREE by 12% on day 1 in the traumatised dogs but on day 2 the PREE only overestimates the MREE by 1.2%. On the other hand, the PREE tended to be below the MREE on both days in the dogs with medical illness by 4.32% on day 1 to 1.4% on day 2 (Appendix 6.9). Inclusion or exclusion of the outlying data points did not have a significant influence on the overall results and so were retained in the analysis reported. Manuscript Menin the format of the American Journal of Veterinary Medicine 3.4 Discussion: During rnetabolism, Iipids, carbohydrates and proteins are bumed in order to provide energy. The potential chernical energy in these molecules is liberated in the body through oxidative processes, which utilises oxygen NO2 ) and subsequently produce carbon dioxide ()/CO2) and water. Dunng the process some of the transferred energy is trapped as ATP and the remainder is lost as heat; however, eventually al1 energy will be converted to heat during a resting tat te.^^^ Calorimetry is a rnethod which measures energy by measuring the rate of heat removal from the body (direct calorimetty) or the rate of heat production by the body (indirect ca~orirnetry)?~~ Indirect calorirnetry is accomplished by measuring the products of oxidative metabolisrn V02 and VC02. Respiratory gas exchange measurements VOz ,VC02, minute ventilation ) and the RQ are detemined frorn the expired gas analy~is?~ The V02 and VC02 rneasurernents are used to determine the REE by the abbreviated ~eirforrnu~a.'~~~~~ Burstein et al" demonstrated in critically il1 humans that there is, on average, a 2% error in the REE when it is determined using only the gas exchange measurernents and discounting the protein metabolisrn. This observation has also been observed in critically il1 dogs admitted to the intensive care unit at the Ontario Veterinary College (Appendices 6.19 and 6.20). The present study dernonstrated that, on average, the means of the MREE and PREE were not statistically different in the medicine and control groups on the two study days. However, the ability, on average, to predict the 24-hour EE on the second day in the postoperative group was poor. If the MREE is considered to be valid the PREE overestimated the energy requirements by 20.95%. Furthemore, there was a trend towards overestirnating the REE by i5.5%, in the traumatised dogs on day 1. Similar findings were reported by Walton et al l6in a study using open-flow calorirnetry to measure REE in 104 post-operative or traumatised dogs (PO&T). When the mean measured REE was compared to the mean of the predictive REE, with an illness factor of 1.6 applied, there was a significant difference between the mean measured REE Manuscript written in the format of the American Journal dveterinary Medicine 91 (40.2 + 3.14 kcaUkg/day) and the mean predicted REE (50 kcalkg/day). In Our study, the combined mean MREE on day 1 (39.9 kcaUkg/day) for the postoperative and trauma groups was similar to Walton et al's findings; however, on day 2 the combined mean was less (37.4 kcal/kg/day). The difference between these two studies may reflect the differences in the frequency of indirect calorimetry readings and analgesic medications and nutritional support were provided in the present study. In the present study MREE was measured three times in a 24-hour period on two days compared to Walton's study which performed only one measurement of the REE on a single day. Several investigator~,~~~'~~have demonstrated that, if the first measurement of the REE was used as an estimate of the 24-hour REE, an overestimation of the nutritional requirements would result. The PREE in Our study for the combined postoperative and trauma groups was considerably less (5.7 kcalldaylkg) than that reported by Walton. This difference was likely due to the range of illness factors 1.2 to 1.5 selected in the current study as compared to the 1.6 illness factor utilised in Walton's study. The mean values only indicate how these two methods MREE and PREE differ on average per day. In Walton's study the mean MREE of the PO&T group was 20% less than the predicted. This is similar to our findings in which the mean MREE was 20% and 15.5% less than the PREE in the postoperative and traumatised dogs on day 1 and day 2 respectively. However, group means do not provide information for determining REE on an individual dog basis, which is more relevant in a clinical setting. It is important to assess how well the MREE and PREE agree in an individual and not just in a group of dogs with a partîcular illnes~!~An acceptable level of clinical agreement between the MREE and PREE had been arbitrarily chosen prior to the start of the study. A difference of 20% or greater between these two methods was deemed to have clinical significance. In the present study, the differences between the methods of REE determination on day 1 ranged from the MREE being greater than PREE by 99% to the MREE being less than PREE by 62% and these differences were inconsistent in al1 the groups on either day. A consistent bias was not observed and, Manuscript htten in the fmat of the Amencan Journal afvetefinary Medicine 92 therefore, a simple correction factor cannot be utilised to account for the difference between the two methods of REE determination. In fact, the two methods agreed (measurements of the REE were between 2 20%) for only 51 % of the dogs on day 1 and 57 % on day 2 with no group demonstrating an acceptable level of agreement on either day. Moreover, there was a wide variation in the REE determinations. The 95 % confidence interval demonstrates that the REE as determined by the PREE may underestimate the MREE by 82% to 83% or overestimate it by 46% to 50%. Therefore, the actual or measured REE could not be predicted reliably from the predictive equation in this study. Dogs in this study had a wide range of MREE to PREE ratios. The MREE was greater than the PREE in 55% of the dogs and less than the PREE in 45%. The relative percentage difference ranged from 0.05% to 99.9%. Some of the discrepancy between the MREE and PREE is inevitable and may be due to the inherent variability of the measured REE or due to measurement error. The open-circuit calonmetry method is shown to be accurate but it requires meticulous calibration of gas sensors, fiow meters, strict protocols and technique.24t25.27-uA6In the present study approximately 1% of the overâll variation in the MREE could be attributed to the individual perforrning the reading. A more extensive training period in the methodology of the system may have decreased or eliminated this random error in Our study. However, this random error was likely not the significant source of the differences observed between the two methods of REE determination. The closer agreement between the two methods in the control dogs an day 1 indicates that the real variability is due to the individual responses to illness or injury in the clinical study. Some degree of error is inevitable in the collection of metabolic measurements in the clinical setting and is inherent in the equations and assumptions that fom the basis of indirect calorimetry? The measured REE rnust be performed in a metabolic steady state in order for an accurate assessment of the REE to be obtained; otherwise artifact may be introduced into the meas~rement.~A metabolic steady state implies that al1 oxygen Manuscript written in the format of the Amencan Journal of Vetennary Medicine 93 mnsumed and al1 carbon dioxide produced is due to metabolic processes. This assumption holds true if the body is at rest and there are no alterations in the body stores of these two gases. The body gas stores are influenced by any factor which alters the minute ventilation such as pain, anxiety, restlessness, hypoventilation or hype~entilation.~In the present study we attempted to minimise any artifact and standardise the measurement of the REE by: defining a steady tat te^^. performing multiple reading~~~~~", performing the test after a period of rest for at least 30 minutes36v3g,allowing at least 30- minutes after administration of analgesic medication", allowing 30 to 60 minutes afier a diagnostic or therapeutic procedure or physical activity (walks)= and three hours after a rnealm6". In the study reported here the coefficient of variation's acceptable range utilised in defining a steady state was expanded from s%~'*~',as utilised in humans, to 10% in the il1 dogs and 15% in the healthy dogs. The expanded range was adopted due to the variable nature of the conscious dogs' respiratory patterns. There was considerable variation in the MREE, which could not be attributed to a single physiologic parameter assessed. However, the "dayn and "agen variables did constitute approximately 24% of the total variability observed in the MREE. Other factors that may influence the metabolic rate of the dog and contribute to the inherent variability of the MREE are the amount of sedation or analgesia administered, nutrient intake, the clinical condition (semi-starvation decreases the REE and sepsis increases the REE), biologic response to illness and altered body However, the individual variation due to these factors is difficult to quantify and account for, but these factors Iikely contributed to the variability observed in the MREE in this study. As a predictor of the MREE, the predictive equation was able to estimate on average the REE in the medicine and control groups; however, the variation in the MREE is too great for an accurate or reliable estimate of the actual REE in an individual based on the PREE. As there is significant variability obsewed in the MREE of individuals with similar diseases or injury, and the ability to precisely estimate the REE in hospitatised individuals is Manuscript written in the format of the Amefican Journal ofveterinary Medicine 94 ,22-24.a.a"~, it is not surprising that these hnro methods did not agree in this skidy. The predictive equation was derived earlier in the twentieth century by measuring EE, in a range of anirnals from as small as a mouse to as large as an elephant." Accuracy of the regression slope is dependent on the range of weights studied. When the range of weights is narrow, large errors in the prediction of the REE oc~ur.'~The predictive equation's ability to predict REE is based on the assumption that the body weight in kilograms to the power function of 0.75 reflects the metabolically active portion of the body and assumes a normal body comp~sition.~ There is compelling evidence that this assumption is incorrect in healthy dogs.lg*" One can expect in il1 dogs that where the measured weight and body composition may be altered due to disease conditions, pre-hospitalisation catabolism, extracellular volume changes and intravenous fluid resuscitation, greater imprecision of the predictive equation's ability to estimate the 24-hour EE will be observed. Another potential problem with the predictive equation lies in the estimation of the severity of illness or injury correction factors. The illness factor have been derived from clinical human data14.17and are not based on experimental data in dogsl'. They are simply a subjective estimate of the severity of illness or injury in a dog and can Vary with the individual assessor. 60th methods presented in this study have difficulties associated with them in their determinations of the REE. Due to the inherent variability associated with measuring the REE via indirect calorimetry, strict adherence to methodology is required for accurate results. However, when applying the predictive equation in the estimate of the RE€ we ignore the significant biological diversity present in the individuals even when a correction factor is applied and may not be valid in healthy In comparison to the predictive equations, the methodology of indirect calorimetry has been validated in a myriad of studies in humans, where the reproducibility and accuracy of the technique under a variety of situations has been compared to direct calorimetry, methanol burning and mass spectr~metry.~~~'The Manuscript written in the format of ttie American Journal of Veterinary Medicine 95 measurement of REE by indirect calorimetry has been shown to be reliable in healthy dogs? However, a 10 to 15 minute adaptation period to the collection system with serial measures of the RE€, discounting the first reading, was required in order to obtain a reliable measure- Currently, there are no reports in the veterinary fiterature, which evaluate the reliability of indirect calorimetry in il1 dogs. The principles of indirect calorimetry and the theory of measuring REE based on the knowledge of V02 and VCOl and nitrogen excretion have been known for almost a ~entury.~~~*~'However, until recently, the application of these principles have been hampered by cumbersome gas collection methods. With the advent of newer technology and portable machines, indirect calorimetry has become the "clinical gold standard" for the measurement of REE at the bedside of hospitalised humans" In conclusion, this study demonstrates that the level of clinical agreement between the MREE and PREE is poor in the clinical setting. The prediction of the REE by the application of the predictive equation did not agree to within + 20% of the measured REE approximately 45% of the time in il1 dogs. In other words approximately 45% of the critically il1 dogs' energy needs cannot accurately be estirnated if the traditional predictive equation for energy expenditure is utilised. No consistent bias was observed between the two methods studied and we were unable to identify a single specific physiologic variable which could predict the MREE with any precision. The level of disagreement between the two methods studied indicates that the importance of accurately estimating the REE in iil or healthy dogs by the predictive equation is unacceptable. Manuscript wriüen in the fonnat of the American Journal afveterinary Medicine Footnotes a TEEM 100 Metabolic apparatusm.Aerosport Inc, Ann Arbor, MI. Eukanuba Maintenance . lams Company, Dayton, OH Hans Rudolph Inc, Kanas City, MO. d Eukanuba Low Residue , lams Company, Dayton, OH 'SAS lnstitute Inc, SAS/STAT@' Software, Cary, NC Manuscript menin the format of the Arnerican Journal ofveterinary Medicine Table 3.1. The descriptive statistics: age (years), weight (kg), illness factor (illness factor) and respiratory quotient (RQ) of the healthy and il1 dogs on a per group basis in the level of agreement study between the two methods of resting energy expenditure determinations (n=77). The means expressed here have not been adjusted for any covariants. Variable Group Median Unadjusted SD Min Max Mean AG€ 2.28 WGT lllness factor RQ C= healthy staff owned dogs; M= client-owned dogs with a variety of medicai illness excluding neoplasia; PO = client owned dogs recovering from a major surgical procedure; T= client-owned dogs in the post-trauma period. Manuscript menin the fmatof the Amefimn JoumaI ofvetennary Medicine Table 3.2. The gender distribution per group of the 77 dogs enrolled in the level of agreement study between the two method of resting energy expenditure detemination; measured versus predicted. Group Fernale Fernale Male Male Total SP~Y- - Neutered number of dogs Control O 5 O 5 1O Medicine Postoperative Trauma Total Manuscript written in the fomat of the Amencan Journal ofveterinary Medicine Ei Measured Day 1 Predicted Day 1 f3 Measured Day 2 PI Predicted Day 2 T Control Medicine Postoperative Trauma Figure 3.1. The adjusted means of the MREE/kg and PREEIkg and the 95% confidence interval of the MREHkg in dogs with a variety of illnesses Le. healthy dogs, medicaI illness, postoperative pen'od and post-trauma period. The level marked with the single asterisk indicates that there is a significant difference (p=0.0181) from one in the logarithm of the ratio for the postoperative group on day 2. In other words the MREE on average is 20.95% less than the PREE in this group on day 2. Manuscript writîen in the fmat of the American Journal of Veterinary Medicine Figure 3.2. The difference on day 1 between rnethods of REE determinations. Measured resting energy expenditure via indirect calon'metry and the predictive equation in the control group (m). dogs with medical illnesses (*). dogs recovering from major surgery (A) and traumatised dogs (0)are plotted against their means. To show that the indirect calorimetry and the predictive equation agree & 20%). al1 the data points should lie between the dotted lines. The axes (x and y) are in the natural logarithmic scale: to revert to the original scale of kcal/day, an antiIog of the values displayed on the graph must be perforrned. Manuscriptwritten in the format of the Amencan Journal ofveterinary Medicine Figure 3.3. The difference on day 2 between methods of REE determinations. Measured resting energy expenditure via indirect calorimetry and the predictive equation in the control group (1)-dogs with medical illnesses (+).dogs recovering frorn major surgery (A) and traumatised dogs (a)are plotted against their means. To show that the indirect calorimetry and the predictive equation agree (+ 20°h), al1 the data points should lie between the dotted Iines. The axes (x and y) are in the natural Iogarithmic scale: to revert to the original scale of kcallday, an antilog of the values displayed on the graph must be performed. Manuscript menin the format of the Arnerican Journal of Vetennary Medicine Table 3.3. Percentage of the data points, which represent the individual dog's resting energy expenditure (REE) deteminations that are within or outside the clinically acceptable limits (+ 20 %). In order for the two rnethods, rneasured REE and predicted REE, to agree there should be no data points outside the acceptable limits of + 20% difference in kilocalories per day. Number Within limits Outside Number of Within limits Outside of Dogs %Agree Iirnits Dogs %Agree limits Day 1 Day 1 % Do not Day 2 Day 2 % Do not Ag ree Agree Day 1 Day 2 Control 7 72 28 4 50 50 Medicine 27 48 52 24 62 38 Postoperative 15 40 60 8 50 50 Trauma 14 57 43 8 50 50 TOTAL 63 51 49 44 57 43 The predictive equation is only accurate to within + 20% of the measured RE€ beilveen 51 % to 57% of the time on the two days tested. The group in which the predictive equation performs the best is the control group on day 1 where 78% of the data points are within the acceptable lirnits. The group in which the predictive equation performs the worst is in the postoperative group on day 1 where only 40% of the data points are within acceptable limits. In order for the rnethods, measured REE and predicted REE to be interchangeable there should be no data points outside the 2 20% Iimit. Manuscript written in Vie fmatof the American Journal of Veterinary Medicine 3.5 Reference List: Barton RG. Nutrition support in critical illness. NCP 1994; 9:127-l39, Giner M, Laviano A, Meguid MM, Gleason JR. In 1995 a correlation between malnutrition and poor outcome in critically il1 patients still exits. Nutrition 1996; 12~23-29. Cerra FB. Hypermetabolism,organ failure, and metabolic support. Surgery 1987; 10l:l-14. Eichacker PO, Hoffinan WD, Banks SM, Danner RL, Richmond S, Wilson L, MacVittie TJ, Natanson CSerial metabolic cart and hemodynamic estimates of total body oxygen consumption in an awake canine model of human septic shock. Am Rev Respir Dis 1992;145:A790. Fettman MJ. Hyperrnetabolism in illness and injury. Comp Conf Educ Pract Vef 1992; l4:1284-1290. Shaw JHF, Wolfe RR. A conscious septic dog mode1 with hemodynamic responses similar to responses of humans. Surgery 1984; 95553-561. Klein SK, Kinney JM, Jeejeebhoy K, et al. Nutrition support in clinical practice: Review of published data and recommendations for future research directions. J Parenter Enterai Nutr 1997; 21:133-156. Krueger KJ, DiPalma JA. The optimal diet for the critically il1 patient. Curr Opin Crif Care 1 997; 3: 127-1 3 1. Askanazi J, Weissman C, R0senbaum.S H. Nutrition and the respiratory system. Crif Care Med 1982; 10:163-172. Brooks MJ, Melnik G. The refeeding syndrome: An approach to understanding its complications and preventing its occurrence. Phamocotherapy 1995; 15: 713. Sheldon GF, Peterson SR, Sanders R. Hepatic dysfunction during hyperalimentation. Arch Surg 1978; 113:504-508. Marino PL, Finnegan MJ. Nutritional support is not beneficial and can be harmful in critically il1 patients, in Crifical Care Clinics: Controversies in Crifical care Medicine 1986; 667-675, Manuscript wriüen in the format of the American Journal of Veterinafy Medicine Armstrong PJ, Saker KS, Davenport DJ, Remillard RL. Assisted feeding in hospitalized patients: Enteral and parenteraf nutrition, 1n:Hand MS, Thatcher CD, Remillard RL,et al ,eds. Small Animal Clinical Nutrition. MarceIine,Missouri: Walsworth Publishing Company, 2000; 351-400. Remillard RL, Thatcher CD. Critical care, in Vetenanry Cliniçs of Norfh Amenca Small Animal Medicine 1 989; 1287-1306. Kronfeld DS. Protein and energy estimates for hospitalized dogs and cats, in Proceedings. PuRna Int Nutr Sym in assoc East Vet Conf Orlando 1991; 5-1 1. Walton RS, Wingfield WE, Ogilvie GK, Fettman MJ, Matteson VL. Energy expenditure in 104 postoperative and traurnatically injured dogs with indirect calorimetry. J Vet Emerg CM Care 1998; 6:71-79. Long CL, Schaffel hi, Geiger JW, Schiller WR, Blakernore WS. Metabolic response to injury and illness: Estimation of energy and protein needs from indirect calorimetry and nitrogen balance. J Parenfer Enteral Nufr 1979; 3:452- 456. Kleiber M. The tire of life: An introducfion to animal energetics. Huntington, New York : Robert E. Krieger, 1975. Heusner AA. Body rnass, maintenance and basal metabolisrn in dogs. J Nufr 1991; 121:S 8-S17. Darmaun DlMaroit S, Martin P, Dumon Hl Robins R, Darmaun D, Nauiet N, Nguyen P.Assessrnent of the effect of weight loss on energy expenditure by doubly-labeled water method in obese dogs. J Vet lnfem Med 2000;14:237. Brissoulis G, Venkataraman S, Thompson AE. Energy expenditure in critically il1 children. Cn? Care Med 2000; 48:1166-1172. Feurer 1, Crosby LO, Mullen JL. Measured and predicted resting energy expenditure in clinically stable patients. Clin Nutr 1984; 3:27-34. Weissrnan C, Kemper M. Askanazi J, Hyman Al, Kinney JM. Resting rnetabolic rate of the critically il1 patient: Measured verses predicted. Anaesthesiology 1986; 64~673-679. Mann S, Westenskow DR, Houtchens BA. Measured and predicted caloric expenditure in the acutely ill. Crif Care Med 1985; 13:173-177. Manuscript written in the format of the American Journal of Veterinary Medicine Bursztein S, Elwyn DH, Askanazi J, Kinney JM , Kvetan V, Rothkopf MM, Weissman C. Indirect calorirnetry: History and ovewiew. 1n:Bursztein S, Efwyn DH, As kanazi J,et ai ,eds. Energy Metabolism, Indirect Calorimetry, and Nutrition. Baltimore, Maryland: Willams & Wilkins, 1989; 2-23. Blaxter KL. The minimal metabolism. 1n:Blaxter KL ,eds. Energy metabolism in animals and man. Cambridge: Cambridge University Press, 1989; 120-146. Kinney JM. Energy metabolism. 1n:Fischer E ,eds. Surgical Nutrition. Boston: Little Brown and Company, 1983. Bursztein S, Saphar P, Singer P, Elwyn DH. A mathematical analysis of indirect calorimetry measurements in acutely il1 patients. Am J Clin Nutr 1984; 50:227- 230. McClave SA, Snider HL. Use of indirect calorimetry in clinical nutrition. NCP 1992; 7:207-221. Bursztein S, Elwyn DH, Askanazi J, Kinney JM . Theoretical framework of indirect calon'rnetry and energy balance-eds. Energy Metabolism, lndirect Calorimetry and Nurition. Baltimore, Maryland: Williams & Wilkens, 1989; 27-81. Gurnp FE, Martin P, Kinney JM. Oxygen consumption and caloric expenditure in surgical patients. Surg Gynecol Obstetr 1973; 137:499-513. Weir JB. New rnethods for calculating rnetabolic rate with special reference to protein rnetabolism. J AppI Physiol 1949; 109: 1-9. Ogiivie GK, Salrnan MD, Kesel 1, Fettman MJ. Effect of anaesthesia and surgery on energy expenditure determined by indirect calorirnetry in dogs with malignant and non-rnalignant conditions. Am J Vet Res 1996; 57:1321-1325. Walters LM, Ogilvie GK, Salrnan MD, Fettman MJ , Joy J, Hand MS, Wheeler SL. Repeatability of energy expenditure measurement in clinically normal dogs by use of indirect calorimetry. Am J Vet Res 1993; 541881-1885. Brandi LS, Bertolini R, Calafa M. Indirect calorimetry in critically iil patients: Clinical applications and practical advice . Nutrition 1997; l3:349-358. Darnask MC, Askanazi J, Weissman C, Elwyn DH, Kinney JM. Artifacts in the measurement of resting energy expenditure. Crit Care Med 1983; 1 1:750-752. Leff ML, Hill JO, Yates AA, Cotsonis GA, Heymsfield SB. Resting metabolic rate: Measurement reliability. J Parenter Enteral Nufr 1987; 11 :354-359. Manusuipt wntten in the format of tfie Amencan Journal ofvetennary Medicine VanLanschot JJB, Feenstra BWA, Vermeiji CG, et al, Accuracy of intermittent metabolic gas exchange recordings extrapolated for diumal variation. Crit Care Med 1988; 16:737. Weissrnan Cl Kemper M, Damask MC, et al. Effect of routine intensive care interactions on metabolic rate. Chest 1984: 86:815-818. Canadian Council on Animal Care. Guide to the care and use of experimental animals volumes 1 and 2, in 1984. Operator's Manual, Aerosport Inc. TEEM 100 Metabolic analysis system total energy expenditure measurement. TEEM 100 Metabolic Analysis System Operator's Manual 1993; 147. Sas lnstitute Inc SASfSTAT software. Changes and Enchanments through release 6.12- Cary,NC : SAS Institute, 1997. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods. Lancet 1986; i:307. Ferrannini E. The theoretical basis of indirect calorirnetry: A review. Metabolis 1988; 37:287-301. Westenskow DR, Schipke C.A., Raymond JL, Saffie JR, Becker JM, et al. Calculations of metabolic expenditure and substrate utilization ffom gas exchange rneasurements. J Parenter Enteral Nutr 1988; 12:20-24. Cunningham KF, Aeberhardt LE, Wiggs BR, Phang PT. Appropriate interpretation of indirect calorirnetry for determining energy expenditure of patients in intensive care units. Am J Surg 1994; l67:547-550. Feurer 1, Mullen JL. Bedside measurement of resting energy expenditure and respiratow quotient via indirect calorimetry. NCP 1986; 1:4346. Swinamer DL, et al. Effect of routine administration of analgesia on energy expenditure in critically il1 patient. Chesf 1988; 92:4-20. Rora AM, Shizgal HM. The HarrÏs Benedict equation reevaluated: Resting energy requirements and the body cell mass. Am J Clin Nutr 1984; 40:168-182. Manuscript writîen in the format of the American Journal ofveterinary Medicine 3.5 Addendum for Manuscript 2: Table 3.4. Adjusted mean + 95% Confidence interval for the measured resting energy expenditure per kilogram (MREEIkg) for each of the 2 days where the n= 77. Group Adjusted 95% CI Adjusted 95% CI REEIkg REEIkg Dayl DayZ Control 32.74 21 -8743.61 27.41 14.78-40.05 Medicine 49.27' 43.84-54.7 45.843 40.3-51 -37 Postoperative 40.37~ 32.95-47.8 33.1 85 24.68-41.69 Trauma 39S2 31-3647.63 41.62 32.35-50.9 Total number 142 (61.5%) 89 (38.5%) of readings ' P= 0.009- different from the control dogs on day 1 ; P=0.074- suggestion of a difference from the control group of dogs on day 1 i3P=O.Ol- different from control dogs on day 2; P=0.016- different from the dogs with medical illness on day 1 ; P=0.04- different from the postoperative dogs on day 1. Data are expressed as means with a 95% confidence interval. The adjusted means have taken into account the effects of group, age ,day and a group'day interaction in the determination of the Manuscript written in the fonnat of the American JoumaI of Veterinary Medicine Table 3.5. Adjusted means of the predicted equation for the various groups per kilogram (PREUkg) for each of the 2 days where the n= 77. Group Adjusted Adjusted REEIkg REEIkg Dayl Day2 Control 31 -20 32.30 Medicine 45.99 45.96 Postoperative 41 -90 41 -98 Trauma 46.72 46.80 The adjusted means have taken into account the effects of group, age, day and a group'day interaction in the determination of the means. Manuscript wntten in the format of the American Journal of Veterinary Medicine Cha~ter4: General Conclusions and Future Studies Several important conclusions rnay be drawn from this multipart study of the use of indirect calorimetry in dogs. The measurement of oxygen consumption per kilogram via indirect calorimetry agrees with a traditional clinical standard, closed- circuit spirometry. A reliable and accurate measurement of REE may be obtained in healthy dogs with open-fiow indirect calorimetry. The rneasurement of REE via open- flow indirect calorimetry dernonstrates significant differences amongst groups of il1 dogs. When dogs with medical illness or traumatised dogs are compared to healthy dogs there is a significant difference in the mean MREE. The REE may be calcuIated solely on gas exchange measurements with less than a 3% error in the 24-hour EE. The predictive equation does not agree with the measurement of REE in il1 or healthy dogs on the hodays studied. The measurement of VOdkg by open-flow indirect calorimetry in the dog demonstrates a clinically acceptable level of agreement with a traditional clinical standard, closed-circuit spirometry. In addition, despite the inherent variability in the measurement of REE, the reliability of senal open-flow indirect calorimetry measurernents is good. The reliability of these measurements was only examined in healthy dogs and they required a 10-minute adaptation period to the collection system and a discounting of the first series to obtain a clinically acceptable measurement of the REE. Variance was observed in these readings even in the most reliable series such that an individual reading may be, on average, 7.3 kcal/kg/day above or below the mean. However, despite this variance, open-flow indirect calorirnetry is a clinically acceptable method for the determination of VOz /kg and consequently REE, in healthy dogs and will provide a flexible and feasible method for the determination of the 24-hour EE in a clinical setting. The central goal of this thesis was to compare the clinical standard in veterinary medicine, an estimation of daily EE by the use of a predictive equation: (70 x (BW in multiplied by an illness factor, to the measurement of REE via an open-flow indirect calonmeter. The means of the two methods utilised in the determination of REE were compared to one another in several groups of dogs: a group of healthy dogs or those with medical illnesses or recovering from major surgery or trauma. The standard predictive equation was on average able to predict the measured REE, in the healthy dogs and those with medical illnesses; however, it tended on average to overestirnate the energy needs of dogs in the post-operative or post-trauma periods. The two methods of REE determination, MREE and PREE, cannot be said to agree or disagree on the basis of correlation or similar means. To determine how interchangeable the two methods are for the assessrnent of an individual patient's REE determination, a limits of agreement method of statistical analysis was performed. The lirnit of agreement analysis dernonstrates that there is a large amount of variation between the two methods of REE determination for the majority of the dogs. The REE determinations did not agree in any group on either of the two days tested. Therefore, the estimation of the REE by the predictive equation cannot be utilised to an degree if precision as a measurement of the REE in the clinical setting. No consistent bias was obsewed between these two methods: a simple correction factor could not be utilised to improve the predictive equation's ability to estimate the measured REE. Furthemore, the physiologic variables: age, gender, breed, group, body temperature, heart rate, respiratory rate, carbon dioxide or bicarbonate concentration in venous blood, were assessed. No one specific variable was able to predict the measured REE with any precision. In general practice the majority of animals do not need extensive hospitalisation or nutritional support. Their physiologic functions are sufficiently resilient to be able to compensate and aid in recovery from illness or injury. Therefore, in a large nurnber of hospitalised cases, the estimation of the REE by the predictive equation will provide veterinarians with a baseline for the energy needs of their patient. However, with the current advances in veterinary medical and surgical procedures and the advent of 24-hour intensive care units, more cornplex and critically il1 dogs are being managed, The degree of illness or injury in these animals tends to necessitate prolonged hospitalisation and nutritional support. Critically iIl anirnals are iess tolerant to any error in the estimation of the REE. A mismatching of their caloric intake to energy expenditures may result in an increase in morbidity or mortality. The imprecision in the estimates of the REE by the predictive equation was clinically unacceptable when it was compared to the MREE in several groups of critically il1 dogs in the current study. The estimates of REE by the predictive equation were only correct to within + 20% of the MREE, approximately 45% of the time in any group of ill dogs on either day. Therefore, the clinical utility of the predictive method for the estimation of the daily EE in a il1 dog is poor. As open-flow indirect calorimetry is touted as the "gold standardn for the bedside measurement of REE in human medicine, we recommend that it be utilised in those il1 dogs which require: (1) prolonged nutritional support; (2) prolonged hospitalisation; and (3) in those individuais who demonstrate an unexpected clinical response to therapy or nutritional support. In situations where indirect calorimetry may not be available or is not feasible and the predictive equation is the "best" that can be provided; we strongly emphasize that it only provides a very rough baseline for the energy needs of that patient. Close monitoring of the individual's clinical response to therapy and the nutritional support is very important in order to account for variable individual responses to illness or therapy . Finally, the assumption that al1 ili or injured dogs have a large increase in their REE was not consistent with the findings reported this thesis. When the groups average MREE were compared to one another there was a significant difference amongst the mean with evidence that dogs with a medical illness or traumatised dogs demonstrate, on average, a hypemetabolic response to illness or injury when their MREE is compared to the MREE of the healthy dogs. However, the increase in the 112 REE was not profound. The results present in the clinical study emphasise the significant individual variation in the response to illness, which is clinically important. Technological advances in the area of gas exchange measurements have made it feasible to accurately and precisely measure REE in dogs by a portable indirect calonmeter at the cageside. Various interactions affect the REE, these include: the presence of pain, anxiety, administration of analgesic medication, walks, performing a physical examination, activity in the [CU and time of day. It will be important for future studies in the utilisation of indirect calorimetry in vetennary medicine to define the effects of each specific routine ICU interactions on the REE. So that results may be compared amongst groups of investigators. What is most needed is a consensus regarding what constitutes a resting state and which protocol should be employed for the rneasurement of REE via open-flow indirect calorirnetry. As continuous care becomes more prevalent in veterinary medicine the importance of accurately assessing energy needs becomes increasingly more important in our patients. The results presented here demonstrate that Our current approach will not be able to meet this goal; therefore, we need to either develop "better" and more flexible predictive equations or become more practiced at the measurement of REE in critically il1 patients. Perhaps the more important question to address is, if we can precisely meet an individual's energy expenditures with appropriate nutritional support, does this translate into irnproved clinical outcornes? The use of indirect calotimetry in clinical veterinary medicine will be a useful tool in answering this question. The application of this technique may also provide prognostic information, on an individual, by identifying a patient's response to illness which has been associated with an increased risk of morbidity or mortality. Cha~ter5: Appendices Appendices VII, Vlll XVlX : Mn- male neutered, M-male, Fs- fernale spay, F-female, C or c- healthy control dogs, M or m-dogs with medical illnesses, PO or po-dogs in the postoperative period, T or t-traumatised dogs, ARF- acute renal failure, CNS- central nervous system, DIGdisseminated ÏntrsvascuIar coagulation, DKA-diabetic ketoacidosis, FB- foreign body, GI- gastrointestinal, HBC- hit-by-car, HOD- hypertrophie osteodystrophy, IBD- inflammatory bowel disease, IMHA- irnmune- medicated haemolytic anemia, ITP- immune-mediated thrombocytopenia, LMN- lower motor neuron signs, LS- lumbosacral, Preop or preop- preoperative period, Postop or postop- postoperative penod, SLE- systemic lupus erythematous, aussie- Australian Shepard, bullmast- bullmasti cocker-American Cocker Spaniel, dane- Great Dane, dobbie-Doberman, golden-Golden Retriever, lab-Labrador Retriever, oldenglish-Old English Sheep Dog, pomeran-Pomeranian, staffie- Staffishire Terrier, schnauzer- Miniature Schnauzer shepard-German Shepard, springer- Springer Spaniel, stpoodle- Standard Poodle, weimaran-Weimaraner. 5.1:Appendix 1. The specifications of the indirect calonmeter (TEEM 100@Aerosport) utilised in the studies reported in this thesis Oxygen Sensor Carbon Dioxide Sensor Type Galvanic Fuel Cell Non-Dispersive lnfrared Range 045% 0-1 0% Accuracy + 0.08% 0.05% Response 0-90% in ~5s 0-90% in <2s Drift Pneumotach Range 0-75 Umin Volume Range 2-30 Umin Accuracy 2 2 of reading or 0.25 Umin Response Time Flow Resistance 8 cm H20@ peak flow 115 5.2: Appendix II, The raw data for the agreement study comparing the 5-minute indirect calorimetry readings (pre-IC and post-IC) of oxygen consumption per kilogram (VOdkg) and the 4minute water filler spirometry readings of VOdkg in 6 normal anaesthetised research dogs. DOG rlME PRE-IC 2"d POST-lC znd Spirometer Spriometer PRE-lC POST-lC 1 2 1 Mean 2 Mean 3 Mean 4 Mean 5 Mean 6 Mean 5.3: Appendix Ill. The measured resting energy expenditure determination (MREE kcal fday and standard error of the mean (SEM) four series and on two consecutive days in the seven staff-owned dogs. These are the best linear unbiased estimates (BLUPS) of the REE. The BLUPS are the best estirnators of the true REE value when the variance components have been taken into account. DOG WEIGHT SERIES MREE SEM MREE SEM Dayl Day 1 Day 2 Day2 5.4: Appendix IV. The oxygen consumption determination NO2mlhin) and standard error of the mean (SEM) for each dog in the reliability study over the four series and on two consecutive days in the seven staff-owned dogs. These are the best linear unbiased estirnates (BLUPS) of the V02. The BLUPS are the best estimators of the true VOz value when the variance wmponents have been taken into acwunt. DOG SERIES V02 SEM V02 SEM Day 1 Day 1 Day 2 Day 2 1 1 136.90 10.47 136.85 10.46 5.5: Appendix V. Estirnate of the variance components associated with the oxygen consumption 0/02)measurernents over the two days. The data in the estimate column was used to compute the interclass correlation coefficients for the V02 over the four sen'es. Level Estimate Standard Probability Value Error DOG 275.25 290.46 0.1 8 DOG*DAY 275.35 184.98 9.90E-04 DOG'SERIES 117.63 91 -66 0.05 DAY*SERIES 1-22 28.46 0.24 DOG*DAY*SERIES 21 7.74 95.54 5.00E-1 O Residual 294.62 24.95 5.6: Appendix VI. The oxygen consumption per kilogram 0(02ml/min/kg) averaged over the four series in each normal staff-owned dog (n=7)on Day 1 and Day 2. These means have not been adjusted for the covariants: age, group day, group'day. DOG Mean Median Minimum Maximum Day 1 1 4.74 4.74 4.74 4.74 2 4.43 4.43 4.43 4.43 3 5.85 5.85 5.85 5.85 4 5.98 5.98 5.97 5.98 5 5.01 5.01 5.01 5.01 6 6.40 6.40 6.39 6.40 7 4.1 5 4.1 5 4.1 5 4.1 5 Day 2 1 4.74 4.74 4.74 4.74 2 4.43 4.43 4.43 4.43 3 5.85 5.85 5.85 5.85 4 5.97 5.98 5.97 5.98 5 5.01 5.01 5.01 5.01 6 6.40 6.40 6.39 6.40 7 4.1 5 4.1 5 4.1 5 4.1 5 120 5.7: Appendix VII. Signalment and underlying disease for the dogs included in the study of agreement between the measured REE via indirect calonmeter and the REE via the predictive equation in the 77 dogs. DOG -AGE GENDER BREED GROUP Angus Border Collie Healthy staff and student owned dog Crockett Labrador Healthy staff and student owned Retriever dog cuny Mixed Healthy staff and student owned dog Duyl Mixed Healthy staff and student owned dog Gloon Mixed Healthy staff and student owned dog Macintoçh Bull Temer Healthy staff and student owned dog MacTavish Mïxed Healthy staff and student owned dog Naia Labrador Healthy stafi and student owned Retriever dog Olivia Standard Healthy staff and student owned Poodle dog Picard Greyhound Healthy staff and student owned dog Abbey Golden Abdominal sepsis, diarrhea Retriever Alfie Miniature IMHA. thrornboembolic disease to Schnauzer hindlimts Angus Great Dane Acute Iiver Fàilure, Ieptoçpimis Asta Fox Temer ITP,GI haemorrhage. neutropenia. vomiting. diarrhea Belle Mixed Pleural effusion, heart murmur, hypoproteinemia, renal disease Brady Miniature ITP Schnauzer Bryn Irish RenaI failure Wolfhound Cajun Labrador Megaesophagus, aspiration Retriever pneurnonia, arrhythrnia Calvin Mixed Pancreatitis Chatiey Miniature Leptospirosis, pancreatitis, ARF Schnauzer Chase Weimaraner Polyarthritis and HOD Duke Labrador Coagulopathy, thrombocytopenia, Retriever increased herand pancreatic enzymes George Springer ITP,Gl haernomhage Spaniel DOG AGE GENDER BREED Goose Labrador retriever Grace Labrador Rodentocide intoxication h retriever Hattie 5yr Arnerïcan lntravascular IMHA Cocker Spaniel lndy 7yr Staffordshire Myasthenia gravis TeMer lndy W 4v Dalmation Icterus and hepatic necrosis, leptoçpirosis Jessie 5yr Spnnger Systemic blastornymsis, CNS & Spanie[ pulmonary involvement Kelsey 8yr Mixed Polydispia. polyuric anorexia. icterus LUCY 8mon Staffordshire Rodentocide intoxication-pleural Tem-er effision Mack 2.5yr Airedale Leucopenia, thrornbocytopenia, immune-mediated disease Maddison 6yr Mixed Coon hound paralysis Marcus 1Oyr IMHA. vorniting Mamie Lethargic, anorexia, skin disease, kbrile Oher 1Oyr Miniature DKA, pancreatitis Poodle ozzy 3yr Australian Sheep dog Panda 3yr Old English Eosinophilic IBD, vomiting and Sheep dog dianhea Russell 3yr Jack Russell Rodentocide toxicity-pleural Terrier effision and retroperitoneal haemorrhage Sadie 9yr Mixed Leptospirosis, pancreatitis, renal insufficiency Surley 5yr Amencan IMHA Cocker Spaniel Thunder 4yr Gennan Shepherd Winston 6yr Labrador Neutrophilic pneumonia Retriever Za ke 5yr Staffordshire Menigioencephalitis Terrier Emma 8yr Greyhound Menigioencephalitis Vannah Boxer Poçt-respiratory arrest , seizuring. aspiration pneumonia Baron Golden Postop linear FB and cirrhotic liver Retriever C.J. Dalrnation Penetrating thoracic FB-postop lung lobectomy DOG GENDER BREED GROUP DlAGNOSlS Cobi Labrador Postop pancreatic abscess Retriever drainage ,neutropenia Bordeaux SL Poodle Gastric dilation and volvulus Cognac Golden SE, post-op iinear FB, Retriever anastomosis and resection Delight Weimaraner Postop cervical ventral slot and stabilisation Kncket Golden Septic arthntis-postop Retriever Mitzie Mixed Ruptured gall bladder, open abdomen Mocha Doberman Postop linear FB Pincher Paddy German lnfarcted colon, intrabdominal Shepherd abscess, pancreatitis Park St Poodle Postop nephrectomy, pancreatitis. melena, anemia, thrombocytopenia Salsa Postop jejunal FB. resection anastomosis, DIC Scout Labrador Migrating FB, abdominal sepsis, Retriever closed abdomen Tali I Golden Postop Iinear FB, anastomoçis and Retriever resection Thor Great Dane Gastric dilation and voIvulus Emma Grey hound Postop ventral slot Vannah 10mon F Boxer Postop intraheaptic shunt, Benson 5yr M Miniature Dog attack +top severe neck Schnauzer trauma ,febrite Bonnie 3yr Fs Geman Postop coxofernoral luxation & Shepherd pelvic fracture Brady ZY~ F Mixed HBC-Postop femoral fracture and tibia1 fracture Buster lyr Mn Boxer HBC, Preop fernoral fracture, scabies, previous encephalitis Qrky 5yr Mn Mixed HBC-Postop coxofemoral8 elbow luxation, stifle joint infection Jagar 1.5yr M Geman HBC, Preop, tetraparalysis. Shepherd ce~icallesion Jessie 3yr Fs Siberian HBC-Postop femoral fracture, Husky pneumothorax, chest tube, Major lyr Mn Mixed HBC-Preop femoral fracture hemorrhagic shock Max 7yr Mn Shetland HBC-Postop pelvic fracture repair, Sheep dog head trauma, Pseudomnas cystitis Penny 8yr Fs Mixed HBGPreop avulsion of skin along the ventrum, febrile Sadie 5yr Fs Pomeranian Dog attack-Postop bite wound closure and sepsis DOG -AG€ GENDER BREEb WElGHT GROUP O Maggie 5yr Fs St Pwdle 21 -7 T HBGPreop fracture skull, spinal luxation and LMN signs to hind Iimbs Sam a'r FS Mixed 22 T HBC- Postop diaphragmatic hemia repair, pancreatitiç Sterling Sr M Weimaraner 29 T Postop pelvic repair and surgical wound infection Thunder 7yr Fs Mixed 17 T HBGPostop LS & coxofemoral luxation & tail amputation Tigger 2~r Mixed 33.4 T HBGPostop coxofemoral luxation and partial splenectomy 5.8: Appendix VIII- The VO, (ml/min), VC02(mumin), 05 NO2/ kg, ml/min/kg), VE (Umin) RQ, (VC02,V02), the calculated 24-hour REE (kcal/day) via the ab breviated Weir formula and the predictive formula 24-hour REE (kcal/day) for the il1 and healthy dogs over the three measurement times on the 2 days for the 77 dogs. Please note that if the respiratory quotient (RQ) was not within the physiologic range (0.67-1) the data was excluded. Age is expressed in years. Time represents the different measurement penods 1; 11 :00-14:OO day 1.2; 16:OO-20:OO day 1.3; 3:OO-6:00 day 1. 4; 11 :00-14:OO day 2.5; 16:OO-20:OO day 2. 6; 3:OO-6:00 day 2. WGT= body weight in kilograms; IER =illness factor. GP AGE SU( BREED TlME V05 VC05 RQ5 05 VE5 MREE PREE WGT IER c 9 fs Mixed 1 108 80 0.7 4.8 5.5 739.47 732.78 22.90 1 c 9 fs Mixed 4 130 84 0.7 5.7 8.5 870.62 732.78 22.90 1 c 4 fs Mixed 1 110 76 0.7 4.9 4.6 744.48 749.52 23.60 1 c 4 fs Mixed 4 110 74 0.7 4.9 3.9 741.31 749.52 23.60 1 c 4 fs Mixed 6 112 72 0.7 4 3.7 749.49 749.52 23.60 1 c 4 mn Bullterr 1 180 144 0.8 6.4 7.7 1249.34 854.33 28.10 1 c 4 mn Bullterr 2 144 104 0.7 4.7 6.1 981.73 854.33 28.10 1 c 4 mn Bullterr 3 132 100 0.8 4.3 5.2 907.32 854.33 28.10 1 c 4 mn BuIIterr 4 98 68 0.7 3.1 3.7 663.72 854.33 28.10 c 4 mn Bullterr 5 60 40 0.7 1.9 2.4 403.78 854.33 28.10 1 c 4 mn Bullterr 6 60 42 0.7 1.9 5.6 406.94 854.33 28.10 1 c 3 mn Lab 2 166 110 0.7 5.8 9.8 1116.06 863.44 28.50 1 c 2 fs Stpoodle 3 90 68 0.8 4.3 4 618.34 689.14 21.10 1 c 8 mn Grey 1 190 136 0.7 5.7 6.1 1293.41 974.72 33.50 1 Hound c 8 mn Grey 2 114 84 0.7 3.4 3.8 779.85 974.72 33.50 1 Hound c 8 mn Grey 3 130 94 0.7 3.9 6.1 886.46 974.72 33.50 1 Hound c 8 mn Grey 4 104 72 0.7 3.1 3.4 704.10 974.72 33.50 1 Hound c 8 mn Grey 5 112 74 0.7 3.4 2.9 752.66 974.72 33.50 1 Hound 125 GP AG€ SU( BREED TlME VOS VC05 RQ5 05 VE5 MREE PREE WGT [ER mn Bullterr 84 0.7 mn Lab 102 0.7 mn Mixed 84 0.7 m Weimaran 106 0.83 m Weimaran 98 0.81 m Weimaran 103.33 0.69 m Weirnaran 100 0.88 m Weimaran 90 1.03 mn Lab 194 0.76 mn Lab 190 0.8 mn Lab 220 0.95 mn Lab 172 0.74 rnn Lab 190 0.88 fs Lab 148 0.84 fs Lab 140 0.89 fs Lab 126 0.84 m Airedale 112 0.94 rn -4iredale 68 0.77 m Airedale 88 0.94 rn Airedale 86 0.8 m Airedale 90 0.75 rn Airedale 94 0.75 m Dalrnation 194 0.78 m Dalmation 232 0.76 rn Dalmation 184 0.83 rn Dalmation 152 0.75 rn Dalmation 164 0.78 fs Shepherd 196 0.7 fs Shepherd 216 0.77 fs Shepherd 214 0.78 fs Shepherd 214 0.75 fs Shepherd 244 0.81 fs Cocker 40 0.7 fs Cocker 52 0.8 fs Cocker 40 0.7 fs Cocker 50 0.7 mn Dane 236 O.? rnn Dane 266 0.7 fs Foxterrier 30 0.7 AGE SU< BREED TlME MREE PREE WGT 8 fs Foxtemer 3 545.76 509.22 9.00 8 fs Foxtemer 6 46 1.84 509.22 9.20 8 fs Foxtemer 2 528.08 474.88 8 -20 8 fs Foxtemer 4 201-89 474.88 8 -20 5 fs Mixed 1 476.35 870.64 18.40 5 fs Mixed 2 492.1 9 870 -64 18-40 5 fs Mixed 3 564-77 870.64 18-40 5 fs Mixed 4 548.93 870.64 18.40 5 fs Mixed 5 666.89 870.64 18-40 12 mn Schnauzer 1 476.35 620.67 12.30 12 mn Schnauzer 2 473.1 8 620.67 12.30 12 mn Schnauzer 3 479.52 620.67 12.30 12 mn Schnauzer 5 620.1 8 620.67 11.90 13 mn Mixed 2 620.1 8 834.67 19-20 13 rnn Mixed 3 983.58 834.67 19-20 13 mn Mixed 4 490.87 834.67 19.60 1 m Lab 1 1993.82 1309.46 35.00 1 rn Lab 2 1753.17 1309.46 35.00 1 m Lab 3 1963.24 1309.46 35.00 1 rn Lab 5 2092.55 1309.46 35.00 4 rnn Springer 1 839.75 1002.1 1 24.50 4 mn Springer 3 1009.44 1002-11 24.50 4 mn Springer 4 980.41 1002.1 1 24.50 4 mn Springer 5 868.78 1002.1 1 24.50 5 fs Cocker 1 274.46 480.62 7.60 5 fs Cocker 2 347.04 480.62 7.60 5 fs Cocker 3 440.99 480.62 7.60 5 fs Cocker 4 1049 -82 1278.14 30.70 7 fs Staffie 6 2044.74 1278.14 30.70 4 fs Dalmation 1 660.56 1411.97 38.70 4 fs Dalmation 2 736.30 1411.97 38.70 4 fs Dalmation 3 750.82 1411 -97 38.70 8 fs Mixed 1 964.05 986-73 24.00 8 fs Mixed 2 972.23 986.73 24.00 8 fs Mixed 3 1062.49 986.73 24.00 8 fs Mixed 5 640.5 1 986.73 24.00 8 fs Mixed 6 993.08 986.73 24.00 0.8 fs Staffie 1 959.04 734.81 16.20 0.8 fs Staffie 2 789.35 734.8 1 16.20 127 GP AG€ SU< BREED TlME V05 VC05 RQ5 05 VE5 MREE PREE WGT IER fs Staffie fs Staffie mn Mixed mn Mixed mn Mixed mn Mixed mn Mixed mn Mixed fs Collie fs Collie mn Minpoodle mn Aussie rnn Aussie fs Oldenglish fs Oidenglish fs OldengIish mn Jackrussel mn Jackrussel mn Jackrussel mn Jackrussel fs Mixed fs Mixed fs Mixed fs Cocker fs Cocker fs Cocker fs Cocker fs Cocker mn Lab mn Lab mn Lab mn Staffie mn Staffie mn Staffie fs Stpoodle fs Stpoodle fs Stpoodle fs Golden fs Golden GP AGE SU< BREED TlME VOS VC05 RQ5 05 VE5 MREE PREE WGT IER fs Golden fs Golden fs Golden fs Golden Mn Stpoodle Mn Stpoodle Mn Stpoodle fs Dalmation fs Dalmation fs Dalmation fs Lab fs Lab fs Lab fs Lab fs Lab Mn Golden fs Weimaran fs Weimaran fs Weimaran fs Greyhound fs Greyhound fs Greyhound fs Golden fs Golden fs Golden fs Golden fs Mixed fs Mixed Mn Dobbie Mn Dobbie Mn Dobbie Mn Dobbie Mn Dobbie Mn Shepherd Mn Shepherd fs Bullmast fs Bullmast fs Bullmast fs Bullmast - AG€ SU< BREED TiME MREE PREE WGT IER 3 fs Builmast 5 1699.83 1558.73 37.40 1-4 3 fs Bullmast 6 1l83.10 1558.73 37.40 1-4 4 Mn Dane 3 2075.39 1659.48 48.00 1.3 0.1 f Boxer 1 733.1 3 637.34 13.40 1.3 0.1 f Boxer 2 542.59 637.34 13.40 1.3 0.1 f Boxer 3 779.85 637.34 13.40 1.3 0.1 f Boxer 4 603.82 637.34 13-40 1.3 0.1 f Boxer 5 424.63 637.34 13.40 1.3 5 m Schnauzer 1 419-62 555.22 10.10 1-4 5 rn Schnauzer 2 419.62 555.22 10.10 1-4 5 m Schnauzer 3 290.30 555.22 10.10 1.4 5 m Schnauzer 4 419.62 555.22 10.10 1-4 5 fs Pomeran 2 72.58 3 17-80 4.80 1-4 5 fs Pomeran 3 145.1 5 317.80 4.80 1-4 3 fs Shepherd 1 1606.92 1330.58 34.00 1.35 3 fs Shepherd 3 1156.44 1330.58 34.00 1.35 3 fs Shepherd 4 1623.28 1330.58 34.00 1.35 1 Mn Boxer 2 1080.17 1337.42 36.00 1.3 2 Mn Boxer 3 1269-39 1337.42 36.00 1.3 1.5 m Shepherd 1 1887.72 1487.91 41 -50 1.3 1.5 m Shepherd 2 1635.44 1487.9 1 41.50 1.3 1.5 m Shepherd 3 1491-61 1487.91 41 -50 1.3 1.5 m Shepherd 4 1777.10 1487.9 1 41 -50 1.3 1.5 m Shepherd 5 1675.15 1487.91 41 -50 1.3 3 fs Husky 1 1O23 -96 1045 -99 23.50 1-4 3 fs Husky 2 697.25 1045.99 23.50 1.4 3 fs Husky 3 865.33 1045.99 23.50 1-4 3 fs Husky 4 1113-70 1045.99 23.50 1-4 3 fs Husky 5 9 17.34 1045.99 23.50 1.4 3 fs Husky 6 1301.07 1045.99 23.50 1.4 5 fs Stpoodle 2 1523.00 1256.22 30.00 1.4 1 Mn Mixed 2 1372.32 1309.46 35.00 1.3 1 Mn Mixed 3 683.25 1309.46 35.00 1.3 1 Mn Mixed 4 702.26 1309.46 35.00 1.3 8 fs Mixed 1 605.66 722.08 15-05 1.35 2 fs Mixed 1 1110.38 930.69 22.20 1.3 2 fs Mixed 2 1188.i9 930.69 22.20 1.3 2 fs Mixed 3 828.67 930.69 22.20 1.3 2 fs Mixed 5 807.55 930.69 22.00 1.3 GP AGE SU( BREED TlME VOS VC05 RQ5 05 VE5 MREE PREE WGT [ER t 2 fs Mixed 6 102 72 0-71 4.7 4.1 692.76 930.69 22.00 1.3 t 2 m Weimaran 3 156 108 0.69 5.4 4.6 1056.15 1180.95 29.00 1.35 t 2 m Weirnaran 4 156 116 0-75 5.4 6.3 1068.83 1180.95 29.00 1.35 t 7 fs Mixed 1 96 64 0.7 2.7 1.3 646.04 879.08 17.00 1.5 t 2 Mn Mixed 1 210 138 0.7 6.4 5.4 1410.05 1361.56 33.40 1.4 t 2 Mn Mixed 2 404 264 0.7 12 11.8 2710.31 1361.56 33.40 1.4 t 2 Mn Mixed 3 430 284 0.7 13 11.7 2889.50 1361.56 33.40 1.4 t 2 Mn Mixed 4 380 252 0.7 12.1 14 2555-14 1361.56 31.60 1.4 t 2 Mn Mixed 5 402 262 0.7 12.7 13.2 2695.80 1361.56 31.60 1.4 t 2 Mn Mixed 6 458 310 0.7 14.4 12.9 3089.55 1361.56 31.60 1.4 131 5.9: Appendix IX Cornparison of the measured REE (kcal/day) and predicted REE on an individual dog basis in the four groups on day 1 and day 2. The dotted lines indicate the level of clinical agreement k20%). Each dot represents a single dog's REE determinations. The axes (x and y) are in the natural logarithmic scale: to revert to the original scale of kcdday, an antilog of the values displayed on the graph must be perforrned. Appendix IX : cont' Appendix IX :cont' Appendix IX :cont' Appendix IX : cont' 1 - û.8 - 0.6 m - E - Cu 3 0.4- Y 4.8 -- -1 - Appendix IX :cont' Appendix IX :wnt' Appendix IX :wnf 1 - 0.8 -- 'fi 0.6 - (O 3 (O 5: 0.4 - w cn u 2 02-- 4 E W OÏE U1 K r) a 5 4.2--. b O E 4.4 - 2 Q) !E 0 4.6 - 4.8 -- n -1 - 5.1 0 Appendix X. The adjusted mean of the measured REE (MREE), predicted REE (PREE) and ratio (MREWPREE) of ail indirect calorimetry readings in each dog with an RQ between 0.67-1.3 averaged over the two days, Ail indirect calorimetry readings in each dog over the hivo days have been averaged and the abbreviated Weir formula has been applied to calculate the MREE (kcavday). Group DOG 'MREEkcalIday PREEkcalIday Ratio duffy gloon macintos h naia oiivia picard mctavis h chase goose gracie mack marcus thunder aifie angus asta asta2 belle brady cajun calvin duke george hattie indy indyw kelsey lucy Group DOG MREE kcalfday PREE kcalfday Ratio maddison 1428.54 1202 1.19 mamie 796.2 1 927.95 0.86 oliver 358.39 432.06 0.83 0- 1207.7 969.46 1.25 panda 1153.54 1351.33 0.85 russell lO3O.95 504.04 2.05 sadie 1301-14 714.31 1.82 surfey 579.07 566.97 1 .O2 winston 1232.99 1256.22 0.98 zake 1726.56 92 1.24 1.87 paris 849.26 910.43 0.93 tali 866.67 175.23 0.74 bordeaux 1216.88 948.78 1.28 ci 1063.89 1032.63 1 .O3 cobi 719.03 1059.33 0.68 cognac 1311.9 1 143.09 1.15 delight 1301.24 1030.67 1.26 emma 1072.33 1 166.49 0.92 krickett 906.06 1278.47 0.71 mitzie 947.03 1312.16 0.72 mocha 1 538.65 1410.19 1 .O9 paddy 121 8.99 1730.99 0.70 salsa 1566.95 1558.73 1.O1 thor 2075.39 1659.48 1.25 vannah 616.8 637.34 0.97 ben 387.29 555.22 0.70 sadie 120.96 317.8 0.38 bonnie 1462.21 1330.58 1.10 buster 1 174.78 1337.42 0.88 jagar 1693.4 1487.91 1.14 jessie 986.44 1045.99 0.94 maggie 1523 1256.22 1.21 major 919.28 1309.46 0.70 penny 605.66 722.08 0.84 sam 925.51 930.69 0.99 Group DOG MREEkcalfday PREEkcalfday Ratio T steriing 1062.49 1180.95 0.90 T thunder 646.04 879.08 0.73 T tigger 2558.39 1361.56 1.88 5.1 1: Appendix XI. The measured REE (MREEikg kcal/kgIday) averaged over the two days in al1 dogs whose RQ was between 0.67- 1.3. Group DOG MREW kg Group DOG MREEl kg C Duffy 35.1 5 Paris 41 -43 Gloon Tali Macintosh Bordeaux Naia CJ Olivia Co bi Picard Cognac Mctavish Deightl Chase Emma Goose Kric ket Gracie Mitzie Mack Mocha Marcus Paddy Thunder Salsa AIfie Thor Angus Vannah Asta Benson Asta2 Sadie Belle Bonnie Brady Buster Cajun Jagar Calvin Jessie Duke Maggie George Major Hattie Penny Indy Sam lndyw Steriing Kelsey Thunder Lucy Tigger Maddison ouy Marnie Panda Oliver Russell Winston Sadie Zake Suriey 143 5.12: Appendix XII. Cornparison of the ratio of the rneasured REE (MREE kcallkgfday) to the predicted REE (PREE kcdkgfday) over the 48- hour period. The single asterisk indicates that there is a significant difference between the mean of the MREE to the mean PREE in the trauma group. The PREE overestimates the MREE by 15% on average. C~ntrd Mane ~~ Tram 144 5.1 3: Appendix XIII. The adjusted mean of the measured REE (MREE) and 95% confidence interval and the adjusted mean of the predicted REE. A significant probability value (pc0.05) indicates that there is a significant difference between the mean of the MREE and PREE in kcaukglday for that group over the 48-hour period Group Adjusted 95% CI Ratio Adjusted Probability MREUkg (MREUPREE) PREEfkg value Control 31 -08' 20.93-41.23 0.98 31.70 0.86 Medicine 47.66 42.77-52.55 1.O3 46.1 2 0.54 Postoperative 38.662 31.73-45.59 0.92 41.99 0.27 Trauma 40.01 32.37-47.64 0.85 47.1 1 0.05 1 p=0.0047- significantly different from the medicine group; 'p=0.038- significantly different from medicine group. 145 5.14: Appendix XIV. The level of clinical agreement of al1 groups between the MREE and the PREE when al1 measures were combined over the 48-hour period. The axes (x and y) are in the natural logarithmic scale: to revert to the original scale of kcal/day, an antilog of the values displayed on the graph must be perforrned. 5.15: Appendix XV. The levei of agreement between the MREE and PREE for the individuai groups over the 48-hour period. Comparçion of the MREE and the PREE in healthy dogç WH a 48 hour p&od -1 J Mean of the REE rnethodç (KcaVday) Appendix XV: con't Cornpanion of the MREE with the PREE in dogs with medical iiiness over a 48 hour period 1 - 0.8 - 4 >, 0.6 - CO P g- OA- x" a* g 02------,,,------4 ------E -5------4 -q J Mean of the RE€ methods (kcaliday) Appendix XV: con? Appendix XV: con't 5.16: Appendix XVI. Percentage of data points, which represent the individual dogs REE determinations averaged over al1 readings on the two days. The points outside the acceptable Iimits are 2 20%. In order for the measured REE and predicted REE to be used interchangeably in the clinic there should be no data points outside the 2 20% Iimits. Wihin limits Outside limits Number of dogs %Agree %Do not Agree Control 86 14 7 Medicine 57 43 30 Postoperative 58 42 12 Trauma 67 33 12 5.1 7: Appendix XVII. The average per group of the oxygen consumption (VOz milmin), oxygen consumption per kilogram 0(02mllmin/kg) and RQ in the consecutive 5-minute readings on the two days in the 77 dogs. These are the means which have not taken the following covariants: age,group,day,group*day,in to account. Variable Group Mean Median SD SE Mean Min Max Numberof readings vo2 Control 121.14 112.00 33.45 7.30 60.00 190.00 21.00 Medicine Postoperative Trauma 02 ControI Medicine Postoperative Trauma RQ Control Medicine Postoperative Trauma 152 5.18: Appendix XVIII. The unadjusted mean of the oxygen consurnption per kilograrn -+ SD in each group averaged over al1 the readings on the hrvo days. control medicine postoperative trauma 5.1 9 Appendix XVIX, Signalment and underiying disease for the dogs enrolled in the study comparing the calculation of the REE including protein rnetabolism (Weir formula) and excluding protein metabolism (abbreviated Weir formula), Weir formula REE (kcal/day)= c(3.94 0/02)+1.1 O(COz))X 14401- 2.1 7 (urine nitrogen in grams). Abbreviated Weir formula REE (kcal/day)= [ (3.94 (Von)+? -1 (VCOz)] X1440. Group DOG Diagnosis WEIR URINE Urine nitrogen (kcaUday) (kcaVday) in grams Chase Polyarthritis and HOD 907.4 897.33 6.3 Abbey abdominal sepsis and diarreha Alfie IMHA ml ARF Charley Leptospirosis pancreatitis, ARF Jessie Systemic blastomycosis- pulmanary and CNS Maddison coon hound paralysis Russell rodeticide intoxication-pleuraf effusion Sadie leptospirosis pancreatitis, renal insufficieixy Vannah Post-respiratory arrest, seizuring, aspiration pneumonia Bordeaux Postop GDV CJ Postop lung lobectomy Emma Postop ventral slot and encephalitis Mitzie Open abdomen & ruptured gall bladder Mocha Postop linear FB Brandy HBC-postop femoral fracture Buster HBGpreop femoral fracture, Scabies, encephalitis Corky HBGPostop coxofemoral and elbow luxation & septic stifle joint Jagar HBC-preop tetraparesis Jessie HBC-Postop femoral fracture, pneumothorax Maggie HBC-Preop skull fracture. spinal luxation & LMN signs to hind Iimbs Major HBC-Preop femoral fracture & hemmorraghic shock Penny HBC-avulsion of skin along ventnim, febnle(preop) Sam HBC-postop diaphragmatic hemia repair, pancreatitis Group DOG Diagnosis WElR URINE Urine nitrogen (kcaUday) (kcaUday) in grams T Sterling HBC-postop pelvic repair & 1096.26 1088.71 3.48 wound infection T Tigger HBC-postop partial splenectomy 2336.62 2305.05 14.55 &coxofemoral luxaticn 5.20 Appendix XX. The level of clinical agreement between the measured REE including protein metabolism weir formula) compared to the measured REE excluding the protein metabolism (Abbreviated Weir formula) in kcal/day. Cornparsion of Measured RE€ lnduding and Exduding Protein Metabolism -0.005 J Mean in the rnethods of REE detemination (Kcaiiday) The difference between the REE calculated including the protein metabolism (Weir formula) and the REE calculated excluding protein metabolism (Abbreviated Weir formula) is plotted against their means. To show that the 2 methods of calculating the REE agree al1 data points should lie between the solid lines. If the outlier is removed than the 2 rnethods including and excluding protein metabolism demonstrate excellent agreement. There is on average a 1.28% difference + 0.8% between the two methods for calculating RE€ and 95% of the data points are below a 2.5% difference. The largest difference observed was 4.5%. Therefore the REE may be calculated accurately by the use of the gas exchange measurements atone excluding protein metabolisrn. 5.21 Appendix XXI. The raw data of the 77 dogs enrolled in the level of agreement study of the two methods for the detemination of the REE. The signalment, gas exchange measurement, measured REE via abbreviated Weir formula (M REE kcailday), predicted REE (PREE kcallday), vital signs, blood gas analysis and senim venous lactate level (mmolB). TIME= 1- 3 during day 1; 4-5 during day 2; 1 and 4 are at the 11 :00-14:OO period, 2 and 5 are at 16:OO-20:OO period, 3 and 6 are at the 03:OO-06:OO period; V02 = oxygen consumption in the 5-minute steady state in ml/min;VC02= carbon dioxide in 5-min steady state in ml/min; RQ= respiratory quotient in 5-min steady state ( VCO2NO2);O2Ikg= oxygen consumption per kilogram in 5-min steady state in ml/min/kg; VE = minute ventilation in 5-min steady state in Umin; 1 ER= illness factor; TEMP= body temperature; RESP RATE= respiratory rate; the vital signs wsre performed pnor to metabolic measurements; COn= venous serum carbon dioxide concentration in mmHg; LAC= serum lactate ievel in mrnolll; HC03= venous serum bicarbonate concentration in mmol/l. GP DOG AGE SE% BREED TlME V02 VC02 RQ OUKG VE MREE PREE iER WEIGHT TEMP HEART RESP CO2 LAC HC03 RATE RATE borderco borderco borderco lab lab lab mlxed mixed mlxed mlxed mlxed mlxed mlxed mixed mlxed mlxed mlxed mlxed mixed mixed mlxad mlxed mixed mixed bullterr bullterr bullterr bullterr bulllerr OP DOG AGE SU( BREEO TlME V02 VC02 RQ 021KG VE MREE PREE IER WEIGHT TEMP HEART RESP CO2 LAC HC03 RATE RATE Ino 4 mn bullterr nala 3 mn lab nala 3 mn lab nala 3 mn lab 01 i 2 fs stpoodle 011 2 fs stpoodle 011 2 fs stpoodle PIC 8 mn greyhoun PIC 8 mn greyhoun PIC 8 mn greyhoun pfc 8 mn greyhoun PL 8 mn greyhoun PIC 4 mn bullterr tav 4 mn lab tav 3 mn rnlxed tav 3 mn mixed abbey 3 fs golden abbey 3 1s golden abbey 3 fs golden abbey 3 fs golden abbey 3 fs golden abbey 3 fs golden aifie 5 fs cocker aMe 5 fs cocker alfie 5 fs cocker aMe 5 fs cocker alfie 5 fs cocker angus 2 rnn dane angus 2 mn dane angus 2 mn dane angus 2 mn dane angus 2 mn dane GP DOG AGE SU( BREED TlME V02 VC02 RQ 0üKG VE MREE PREE IER WElGHT TEMP HEART RESP CO2 LAC HC03 RATE RATE asia foxterri asia foxterri asta foxterri asta fo~terri asta2 foxterri asta2 foxterri asta2 foxterri asb2 foxteni asta2 foxterri belle mixed belle mlxed belle mixed belle mixed belle mlxed brady schnauze brady schnauze bmdy schnauze brady schnauze brady schnauze bryn wolfioun bvn wolfhoun brYn wolfhoun wolihoun cajun lab cajun lab cajun lab cajun lab calvin mlxed calvin mlxed calvin mlxed calvin mlxed GP DOG AGE SEX BREED TlME V02 VC02 RQ 02/KG VE MREE PREE IER WEIGHT TEMP Hf3RT RESP CO2 LAC HC03 RATE RATE M Ch 0.5 m welmaran 1 128 106 0.83 7.6 8 894.12 703.26 1.2 17,OO 38,7 140 24 34.30 1.90 16.70 Ch welmaran Ch welmaran Ch welmaran Ch welmaran charley schnauze chariey schnauze chariey schnauze chariey schnauze duk lab duk lab duk lab duk lab duk lab emm greyhoun emm greyhoun emm greyhoun emm greyhoun geor springer geor springer gew springer W'Jf springer springer gew springer Gos lab Gos lab Gos lab Gos lab Gos lab Gos lab Gra lab Gra lab GP DOG AGE SU( BREED TlME V02 VC02 RQ 02/KG VE MREE PREE IER WEIGHT TEMP HEART RESP CO2 LAC HC03 RATE RATE m Gra 0.9 fs lab m Gra 0.9 1s lab m haiüe 5 fs cocker m hatlie 5 fs cocker m hattle 5 fs cocker m haiüe 5 fs cocker m hattle 5 fs cocker rn Indy 7 fs staffie m indy 7 fs staffie m indy 7 fs siaflie m Indy 5 fs cocker m indy 7 1s staffle m indy 7 fs stafiie m lndyw 4 fs dalmaîio m lndyw 4 fs dalmaîio m indyw 4 fs dalmaîio m jessle 5 fs springer m jessle 5 fs springer m jessle 5 fs springer m jessie 5 fs springer m kelsey 8 fs mlxed m kelsey 8 fs mlxed m kelsey 8 fs mixed m kelsey 8 fs mixed m kelsey 8 fs mlxed m lucy 0,8 fs staffle m lucy 0,8 fs staftie m lucy 08 fs staffie m lucy 0.8 fs staMe m Mac 2.5 m airdale m Mac 23 m airdale m Mac 2.5 m alrdale GP DOG AGE SU( BREED TlME V02 VC02 RU OUKG VE MREE PREE IER WEIGHT TEMP HEART RESP CO2 LAC HC03 RATE RATE m Mac 2.5 m airdale 4 m Mac airdale m Mac alrdale m rnad mlxed m rnad mlxed m rnad mixed rn rnad mlued rn rnad mlxed m rnad mlxed m Mar datmaUo m Mar dalmaUo m Mar dalmatlo m Mar dalmaUo m Mar dalmaUo m Mar dalmaUo m mam collie m mam collle m mam collle m ollver mlnpoodl m ollver mlnpoodl m ollver mlnpoodl m oliver minpoodl m oliver mlnpoodl m allver mlnpoodl m Oq aussle m ony aussle m Ony aussle m ony aussle m ony aussle m ony aussle m panda oldengll m panda oldengll GP DOG AGE SEX BREED TlME V02 VC02 RQ OUKG VE MREE PREE IER WEIGHT TEMP HEART RESP CO2 LAC HC03 RATE RATE m panda 3 fs oldengli 3 180 130 0.7 5 13.6 1227.17 1351,33 1.3 36.50 39,s 66 80 m panda 3 fs oldengll m panda 3 fs oldengli m panda 3 fs oldengll m mssell 3 mn jackniss m nissell 3 mn jackniss m russell 3 mn jackruss m russell 3 rnn jackniss m nissell 3 mn jackniss m sadle 9 fs mlxed m sadle 9 fs mlxed m sadle 9 fs mlxed m sadie 9 fs mlxed m sadle 9 fs mlxed m sadle 9 fs mlxed m surley 5 fs cocker m sutiey 5 fs cocker m sutiey 5 fs cocker m surley 5 fs cocker m surley 5 fs cocker m sutiey 5 fs cocker m Thu 4 fs shepard m Thu 4 fs shepard m Thu 4 fs shepard m Thu 4 fs shepard m Thu 4 fs shepard m Thu 4 fs shepard m vannah O,1 f boxer m vannah 0.1 f boxer m vannah 0.1 f boxer m vannah 0.1 f boxer m winston 6 rnn lab GP DOG AGE SU( BREED TlME V02 VC02 RQ 02/KG VE MREE PREE IER WEIGHT TEMP HEART RESP CO2 LAC HC03 RATE RATE m winston 6 mn lab 2 254 138 0.5 8,5 14.5 1659,69 1256.22 1,4 30,OO 38.7 m wlnston lab m wlnston lab m winston lab m zake stafiie m zake stafiie m zake stame m zake staMe po baron golden po baron golden po baron golden po baron golden Po bord slpoodle po bord stpoodle po bord slpoodle Po bord stpoodle po bord stpoodle Po bord stpoodle PO cj dalmaîio PO CJ dalmaîio PO cl dalmatio PO cj dalmaîio PO cj dalmaîio po cobl lab po cobl lab po cobl lab po cobl lab po wbi lab po cobl lab po cognac golden po cognac golden po cognac golden GP DOG AGE SEX BREED TlME V02 VC02 RQ 02NG VE MREE PREE IER WEIGHT TEMP HEART RESP CO2 LAC HC03 RATE RATE po cognac mn golden po cognac mn golden po cognac mn golden po del fs welmaran po del fs welmaran po del fs welmaran po emrn fs greyhoun po emm fs greyhoun po emrn fs greyhoun po emm fs greyhoun po emm fs greyhoun po kdc fs golden po krlc fs golden po kdc fs golden po kric fs golden po mlhie fs mixed po mihle fs mlxed po mltzle fs mlxed po mocha mn dobbie po mocha mn dobble po mocha mn dobble po mocha mn dobble po mocha mn dobble po mmha mn dobble Po paddy mn shepard po paddy mn shepard Co paddy mn shepard PO paddy mn shepard po paddy mn shepard po Par fs stpoodle po Par fs stpoodle po Par fs stpoodle GP DOG AGE SEX BREED TlME V02 VC02 RQ 021KG VE MREE PREE IER WEIGHT TEMP HEART RESP CO2 LAC HC03 RATE RATE salsa bullmast salsa bullmast salsa bullmast salsa bullmast salsa bullmast salsa bullmast scout lab scout lab scout lab scout lab Tali golden Tall golden Tal t golden TalI golden Tall golden Tall golden thor dane aiw dane vannah boxer vannah boxer vannah boxer vannah boxer vannah boxer Ben schnauze Ben schnauze Ben schnauze Ben schnauze bonn shepard bonn shepard bonn shepard bonn shepard bonn shepard GP DOG AG€ SEX BREED TlME V02 VC02 RQ 02MG VE MREE PREE IER WEtGHT TEMP HEART RESP CO2 LAC HC03 RATE RATE t bonn 3 fs shepard 6 116 70 0.6 3.4 4.3 769.02 1330,58 1,4 34.00 38.8 88 24 brandy 2 fs mlxed brandy 2 fs mlxed brandy 2 fs mined brandy 2 fs mixed brandy 2 fs mlxed buster 1 mn boxer buster 2 mn boxer W~Y5 mn mlxed W~Y5 mn mixed WrkY 5 mn mixed CO~Y 5 mn mlxed WkY 5 mn mixed jas 1.5 m shepard la9 1.5 m shepard la9 1.5 m shepard las 1.5 m shepard Jas 1.5 m shepard jessie 3 fs husky jessie 3 fs husky jessie 3 fs husky jessfe 3 fs husky jessie 3 fs husky jessie 3 fs husky maggie 5 fs stpoodle maggie 5 fs stpoodle maggie 5 fs stpooâle major 1 mn mixed major 1 mn mixed major 1 mn mixed major 1 rnn mined major 1 mn mined GP DOG AGE SU BREED TiME V02 VC02 RQ 02/KG VE MREE PREE /ER WEiGHT TEMP HEART RESP CO2 LAC HC03 RATE RATE T major mn mixed 6 T rnax mn sheltie T max mn sheltie T max mn sheliie T max mn shelUe T rnax mn sheltie T max rnn sheltie T penn fs mixed T penn fs mixed T penn fs mixed T penn fs mixed T Sad fs pornerani T Sad fs pomeranl T Sad fs pomeranl T sam fs mixed T sam fs mixed T sam fs mlxed T sam fs mixed T sam fs mlxed T sam fs mixed T ster m weimaran T ster m welmaran T ster m weimaran T ster m weimaran T ster m weimaran T thunder fs mixed T thunder fs mlxed T thunder fs rnixed T tigger mn mlxed T Ugger mn mixed T tigger mn mixed T Ugger mn mixed GP DOG AGE SEX BREED T1Mf V02 VC02 RQ 021KG VE MREE PREE [ER WEIGHT TEMP HEART RESP CO2 LAC HC03 RATE RATE T Ugger 2 mn mlxed 5 402 262 0.7 12.7 13.2 2695.8 1361.56 1,4 31.60 38.2 116 24 T tig~er 2 mn mixed 6 458 310 0.7 14.4 12.9 3089.55 1361.56 1.4 31.60 38.2 732 24