Comparison of direct and Doppler arterial blood pressure measurements in rabbits during isoflurane anaesthesia

Veterinary Anaesthesia and Analgesia, 2012, 39, 174–184

doi:10.1111/j.1467-2995.2011.00685.x

RESEARCH PAPER

Comparison of direct and Doppler arterial blood pressure measurements in rabbits during isoflurane anaesthesia Louise Harvey, Toby Knowles & Pamela J Murison School of Clinical Veterinary Science, University of Bristol, Langford House, Langford, Bristol, UK

Correspondence: Louise Harvey, School of Clinical Veterinary Science, University of Bristol, Langford House, Langford, Bristol BS40 5DU, UK. E-mail: [email protected]

Abstract Objective To measure the level of agreement between Doppler measured (DOP) arterial blood pressure (ABP) in the forelimb and directly measured (DIR) auricular systolic ABP (SAP) and mean ABP (MAP) in isoflurane-anaesthetized rabbits. Study design Prospective clinical study. Animals Data were analysed from 17 of 24 healthy rabbits, weighing 1.3–2.8 kg. Methods Rabbits were anaesthetized for neutering using a standardized protocol. A 26G catheter placed in an auricular artery was connected via heparinised saline filled non-compliant tubing (regularly flushed) to a calibrated pressure transducer (zeroed level with the thoracic inlet) to obtain DIR ABP. A cuff was placed proximal to the carpus (approximately level with the thoracic inlet) and a Doppler transducer sited over the dorsal carpal branch of the radial artery to obtain DOP ABP. Simultaneous DIR and DOP ABP recordings were made every 5–10 minutes during anaesthesia. Agreement was assessed as described by Bland JM & Altman (2007). Results Mean ± SD cuff width: limb circumference ratio was 0.50 ± 0.04. Mean between-method differences ± SD, DIR SAP- DOP and DIR MAP- DOP, were +1 ± 8 and )13 ± 8 mmHg respectively. The 95% limits of agreement between DIR SAP and DOP and between DIR MAP and DOP were )14 to +17 and 174

)28 to +2 mmHg respectively. Differences between DIR SAP and DOP were £10 mmHg 85% of the time. Defining hypotension as either DIR SAP < 80 mmHg or DIR MAP < 60 mmHg, and taking DOP ABP of <80 mmHg to indicate hypotension, sensitivity and specificity were 92% and 67% respectively. Conclusions Good agreement was found between DIR SAP and DOP. Doppler measurements below 80 mmHg are a reliable indicator of arterial hypotension. Clinical relevance DOP is acceptable for monitoring ABP in isoflurane-anaesthetized rabbits and is useful for detection of hypotension. Keywords anaesthesia, arterial Doppler, rabbit.

blood

pressure,

Introduction Hypotension may occur during anaesthesia of rabbits using inhalation agents (Imai et al. 1999). It may occur as a result of cardiovascular depression caused by anaesthetic (Branson 2007; Steffey & Mama 2007) and sedative drugs (Lemke 2007) and may be exacerbated by surgical haemorrhage. It may compromise the perfusion of vital organs. Measurement of arterial blood pressure is now considered a basic standard in human anaesthesia (Association of Anaesthetists of Great Britain and Ireland, 2007) and is increasingly performed during anaesthesia of animals (American College of Veterinary Anesthesiology, 1995).

Doppler blood pressure in anaesthetized rabbits L Harvey et al.

Direct measurement of ABP is the accepted standard (Ward & Langton 2007). It provides continuous measurement, pressure waveform display and access for arterial blood sampling for blood gas analysis. Disadvantages are the need for technical skill for arterial catheterisation (especially very small patients), risk of infection, thromboembolism, haemorrhage, haematoma, problems maintaining function of the implanted materials, accidental intra-arterial drug injection and expense of equipment (Wagner & Brodbelt 1997; Love & Harvey 2006). Indirect methods of arterial pressure measurement are non-invasive and generally require less expertise, but are less accurate. Most indirect methods of ABP measurement require a superficial artery on a distal extremity around which an occlusive cuff which is wrapped. The cuff is inflated to a pressure above SAP which temporarily occludes arterial blood flow and then deflated gradually. In the Doppler technique, arterial blood flow is detected by placing a Doppler transducer over an artery distal to the cuff producing an audible signal. The pressure within the cuff when arterial blood flow is first heard returning during deflation of the cuff should be close to SAP (Wagner & Brodbelt 1997). Few studies compare direct and indirect ABP measurement techniques in rabbits. Many of these studies found good correlation between methods studied but the indirect techniques require equipment which is not widely available or easily applied to anaesthetized rabbits (Morizono et al. 1975; Wilson et al. 1975; Newton et al. 1991; Herrold et al. 1992; Kurashina et al. 1994). More recently, the oscillometric technique has been compared to direct measurement of ABP in the abdominal aorta in rabbits (Ypsilantis et al. 2005). The Doppler technique has been compared to direct measurement of ABP in several species, including the cat, dog and horse (Grandy et al. 1992; Bailey et al. 1994; Binns et al. 1995; Caulkett et al. 1998; Haberman et al. 2006), but not in rabbits. An advantage of the Doppler technique is that pulse detection is rarely a problem (Binns et al. 1995; Caulkett et al. 1998). This could be an important advantage in rabbits as they are often small. Other advantages are the wide availability and moderately low cost of equipment. In a recent clinical study of rabbits undergoing anaesthesia, arterial blood pressure (ABP) was measured using the Doppler technique and readings were frequently low (Martinez et al. 2009). The

accuracy of these measurements was, however, not known. The aim of this study was to assess the level of agreement between forelimb Doppler ABP measurements and direct measurements of auricular SAP and MAP in rabbits during isoflurane anaesthesia. Materials and methods Animals Twenty four, American Society of Anaesthesiologists physical status category 1 (American Society of Anesthesiologists 2011), client-owned rabbits of various breeds (12 male, 12 female, estimated age 12 weeks–36 months, weighing 1.13–2.85 kg) undergoing general anaesthesia for elective neutering were studied. The study was approved by the internal institutional ethics review committee and informed owner consent was obtained. Anaesthesia After pre-anaesthetic physical examination, the hair on the external pinnae of both ears was clipped and a eutectic mixture of lidocaine and prilocaine (EMLA Cream 5%; AstraZeneca, UK) was applied over both the central auricular artery and marginal auricular vein beneath an occlusive dressing, for at least 30 minutes. Pre-anaesthetic medication with fentanyl (30 lg kg)1) and fluanisone (1.5 mg kg)1) (Hypnorm; VetaPharma, UK) was then administered by intramuscular injection into the lumbar musculature. Approximately 5 minutes later, a 24 gauge catheter (Hospira Venisystems; Abbocath-T, Ireland) was placed in a marginal auricular vein. Anaesthesia was then induced by intravenous (IV) injection of propofol (Propoflo; Abbott Laboratories, UK, or Propoclear; Pfizer Animal Health, UK) administered to effect. Tracheal intubation was attempted using a blind technique with an uncuffed endotracheal tube (2.5–3.5 mm internal diameter). If tracheal intubation was unsuccessful, anaesthesia was maintained via a mask. Anaesthesia was maintained using isoflurane (Isoflo; Abbott, UK) vaporised in oxygen delivered via a T-piece with Jackson Rees modification using a fresh gas flow of 600 mL kg)1 minute)1. A pulse oximeter probe with a transmittance transducer (Nonin 8500 Series; Nonin Medical, Inc, MN, USA) placed on the tongue was used to measure arterial haemoglobin saturation and to aid measurement of pulse rate. Respiration was monitored by

 2012 The Authors. Veterinary Anaesthesia and Analgesia  2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 174–184

175

Doppler blood pressure in anaesthetized rabbits L Harvey et al.

observing thoracic excursions and the movement of the reservoir bag, and using side-stream capnography (Capnomac Ultima; Datex Engstrom, Finland). Normal saline (Vetivex; Arnolds, UK) was administered IV throughout anaesthesia at a rate of 10 mL kg)1 hour)1. All rabbits received either meloxicam (0.5 mg kg)1) (Metacam; Boehringer Ingelheim, UK) or carprofen (4 mg kg)1) (Rimadyl; Pfizer, UK) by IV injection during surgery. Some rabbits received buprenorphine (0.05 mg kg)1) (Vetergesic; Alstoe, UK) IV intra-operatively. In some rabbits, local anaesthetic solution was infiltrated into the incision upon completion of surgery using either 4 mg kg)1 lidocaine (Lidocaine 1%, Hameln Pharmaceuticals, UK) or 1 mg kg)1 bupivacaine (Marcain Polyamp 0.25%; AstraZeneca, UK). To minimise hypothermia, a warm ambient temperature was maintained and the rabbit was placed on an insulated electrical heat mat. Hypotension was defined as DIR MAP <60 mmHg or DIR SAP <80 mmHg (Gaynor et al. 1999). If hypotension was observed, the vaporiser setting was reduced where possible, the rate of fluid administration increased and/or succinylated gelatin in saline (Gelofusine; Dechra Veterinary Products, UK) administered. In recovery, all rabbits were observed for signs of pain, to allow for administration of further analgesia as required. Direct measurement of ABP After induction of anaesthesia and orotracheal intubation, the rabbit was maintained in sternal recumbency for arterial catheterisation. A 26 gauge catheter of length 19 mm (Hospira Venisystems; Abbocath-T, Ireland) was placed in the middle segment of the central auricular artery of the other ear to the one with the venous catheter. The catheter was connected to a three way tap and flushed with heparinised saline (5 IU mL)1). The ear and catheter were supported by a wad of gauze swabs. The three way tap was connected via heparinised saline filled non-compliant extension tubing 200 cm long with an internal diameter of 1.0 mm (Vygon, France) to a pressure transducer (BD DTXplus; Becton Dickinson Critical Care Systems Pte Ltd, Singapore) connected to an electronic monitor (Kontron Minimom 7138B; Kontron Instruments, UK). The calibration of the transducer was checked against a mercury manometer at the beginning of each study day. Once the rabbit was positioned in dorsal recumbency for surgery, the thoracic inlet was used as the zero level for the 176

transducer. The ears of the rabbit were allowed to lie flat against the mattress on which the rabbit was positioned. In order to minimize the risk of arterial damage, only one attempt was made to catheterise the auricular artery. If arterial catheterisation was unsuccessful on the first attempt, the rabbit was excluded from the study but neutered as intended. A single transducer was used for all the rabbits in this study. Contamination was avoided, using sterile heparin saline for flushing, capping ports with sterile bungs when not in use and using new extension tubing for each rabbit. Doppler measurement of ABP The hair on the medial left carpus was clipped. Acoustic coupling gel was applied liberally to the skin over the dorsal carpal branch of the radial artery (Popesko et al. 2002) which was identified by palpation. An 8.2 MHz frequency ultrasonic Doppler flow detector transducer (Model 811-B; Parks Medical Electronics, OR, USA) was positioned over the artery (ensuring maximal flow signal) and secured in place with adhesive tape. The circumference of the limb immediately proximal to the carpus was measured using a flexible tape-measure. A paediatric blood pressure cuff (StatCorp, FL, USA) with bladder width approximately 40–50% of the limb circumference was then applied at this level and secured with short strips of loosely applied adhesive tape. An aneroid sphygmomanometer (Gold Series DS66 Trigger Aneroid; Welch Allyn, UK) was attached to the blood pressure cuff. The calibration of the sphygmomanometer was checked against a mercury manometer at the beginning of each study day. Headphones (PX-92; Proluxe, China) were used to listen to the Doppler flow signal. Doppler measurements of ABP were taken as follows: The cuff was inflated over 1–2 seconds until flow sounds were no longer audible, then immediately deflated at approximately 2 mmHg second)1. The sphygmomanometer reading at the first reappearance of the flow signal was taken as DOP. The cuff remained completely deflated between measurements. Data collection Pairs of direct and Doppler ABP measurements were recorded every 5–10 minutes during anaesthesia by a single experienced clinician. The numerical display of direct measurements was kept out of sight

 2012 The Authors. Veterinary Anaesthesia and Analgesia  2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 174–184

Doppler blood pressure in anaesthetized rabbits L Harvey et al.

during Doppler measurement and whilst the result was written down. Direct measurements were then recorded immediately (within 1 second of the Doppler flow signal returning) so that pairs of readings were as simultaneous as practically possible. The arterial catheter was gently flushed with 0.3–0.5 mL heparinised saline 15 seconds before commencing cuff inflation. The pressure waveform was subjectively inspected for excessive damping at every measurement. Damping was judged to be excessive if the pressure waveform was obviously flattened and/or if the pulse pressure increased after flushing but then decreased again. If excessive damping was not resolved by flushing and/or slight repositioning to ensure no kinking or compression, all data from that rabbit was excluded from subsequent analysis. At the end of anaesthesia, the catheter in the auricular artery was removed and manual pressure applied for 2 minutes. Rectal temperature was also recorded at this time. The owners of all the rabbits were contacted during the following week and asked about any complications (e.g. swelling) relating to catheterisation of the artery. Statistical analysis Agreement between direct measurements of SAP and MAP and Doppler measurements was assessed as described by Bland & Altman (2007) (for agreement between methods with multiple observations per individual where the true value is non-constant), producing values for the mean difference between measurement methods and standard deviation. Doppler measurements were subtracted from direct ABP measurements. The required univariate analysis of variance was calculated using PASW version 18. A (IBM, UK) test of slope by regression analysis was performed to assess whether or not the mean difference between measurement methods was consistent over the measured range. Betweenrabbit and within-rabbit components of variance for the differences between pairs of directly measured SAP and DOP measurements were also calculated using the components of variance analysis within SPSS version 18. To test the usefulness of Doppler measurements for identifying the presence or absence of hypotension, sensitivity, specificity and predictive values were calculated. Actual hypotension was defined as a direct SAP <80 mmHg and/or a direct MAP <60 mmHg (Gaynor et al. 1999). A Doppler mea-

surement below 80 mmHg was taken as possible hypotension. Sensitivity was calculated as the number of Doppler measurements indicating hypotension divided by the number of direct measurements indicating actual hypotension. Specificity was calculated as the number of Doppler measurements not indicating hypotension divided by the number of direct measurements indicating no actual hypotension. The positive predictive value was calculated as the number of direct measurements indicating actual hypotension divided by the number of Doppler measurements indicating hypotension. The negative predictive value was calculated as the number of direct measurements indicating no actual hypotension divided by the number of Doppler measurements not indicating hypotension. The percentages of Doppler measurements within 5, 10 and 15 mmHg of the direct SAP were also calculated. The differences between directly measured SAP and Doppler measurements were numbered sequentially in the order in which the data were collected, and plotted against the sequence number (see Fig. 3) in order to show changes occurring during the study period. Pearson’s product moment correlation coefficient was calculated to assess the association between direct and Doppler measurements. Data are expressed as mean ± standard deviation. Results The mean dose of propofol used for induction of anaesthesia was 8.1 ± 3.5 mg kg)1. Twenty-three of the 24 rabbits were successfully intubated. Meloxicam was administered to 22 rabbits, carprofen to two rabbits and buprenorphine to 11 rabbits. Two rabbits received incisional lidocaine and two rabbits received incisional bupivacaine. The mean duration of anaesthesia was 78 ± 19 minutes. Five rabbits were excluded due to failure to insert a catheter in the auricular artery. The data from two rabbits were excluded because of excessive unresolvable damping of the direct ABP waveform. A total of 190 paired measurements, collected from the remaining 17 rabbits (11 males, six females, weighing 1.3–2.8 kg), were used for analysis. The numbers of paired measurements collected from the rabbits ranged from 4 to 17. Mean cuff width: limb circumference ratio was 0.50 ± 0.04. Doppler measurements were obtained without difficulty. The mean between-method difference comparing directly measured SAP and

 2012 The Authors. Veterinary Anaesthesia and Analgesia  2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 174–184

177

Doppler blood pressure in anaesthetized rabbits L Harvey et al.

Doppler measurements was +1 ± 8 mmHg. The difference was consistent over the measured pressure range (directly measured SAP range 44– 99 mmHg, directly measured MAP range 33– 83 mmHg) (test of slope by regression analysis : b = )0.043, SE = 0.053, t = )0.817, p = 0.415). The 95% limits of agreement between directly measured SAP and Doppler measurements were )14 to +17 mmHg (Fig. 1). The mean betweenmethod difference comparing directly measured MAP and Doppler measurements was )13 ± 8 mmHg. There was a small but marginally significant difference over the measured pressure range (test of slope by regression analysis: b = )0.108, SE = 0.054, t = )2.013, p = 0.046). The 95% limits of agreement between directly measured MAP and Doppler measurements were )28 to +2 mmHg (Fig. 2). For the differences between directly measured SAP and Doppler measurements, the between-rabbit component of variance was 80% of the total variance and the within-rabbit component of variance was 20% of the total variance. Differences between directly measured SAP and Doppler measurements were £5 mmHg 59% of the time, £10 mmHg 85% of the time, and £15 mmHg

92% of the time. Defining hypotension as either DIR MAP <60 mmHg or DIR SAP <80 mmHg, and taking DOP ABP of <80 mmHg to indicate hypotension, sensitivity was 92% and specificity was 67%. The prevalence of hypotension was 79%, giving a positive predictive value of 91% and a negative predictive value of 68%. The differences between directly measured SAP and Doppler measurements are plotted against the sequence number in Fig. 3. The sample correlation coefficient between directly measured systolic blood pressures and Doppler measurements was +0.82. The sample correlation coefficient between directly measured mean blood pressures and Doppler measurements was +0.79. Hypotension was observed in 15 out of the 17 rabbits at one or more measurement. In six of the 17 rabbits, a directly measured mean blood pressure <45 mmHg occurred at least once. Four of the 17 rabbits received succinylated gelatin in saline in addition to normal saline as intravenous fluid therapy. No rabbits suffered complications due to arterial catheterisation. All rabbits had rectal temperatures >37.0 C at the end of anaesthesia. None of the

30 Upper limit of agreement

SAP – DOP (mmHg)

20

+17.0

10 Mean difference +1.0

0 –10

–14.0

Lower limit of agreement

–20 –30 –40 40.0

50.0

60.0 70.0 80.0 (SAP + DOP)/2 (mmHg)

90.0

100.0

Figure 1 Bland and Altman plot comparing Doppler measurements (DOP) with directly measured systolic arterial pressure (DIR SAP) in 17 rabbits in which anaesthesia was maintained with isoflurane, totalling 190 comparisons. Each data point corresponds to the difference between a pair of Doppler and direct-measurements plotted against the mean of the two simultaneously obtained results. Horizontal lines indicate mean difference and the 95% limits of agreement. The trend line represents the change of mean difference across the measured range. A test of slope by regression analysis showed that slope of this line was not significantly different from 0 (b = )0.043, SE = 0.053, t = )0.817, p = 0.415). 178

 2012 The Authors. Veterinary Anaesthesia and Analgesia  2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 174–184

Doppler blood pressure in anaesthetized rabbits L Harvey et al.

30

MAP – DOP (mmHg)

20 10 Upper limit of agreement +2.0

0

Mean difference

–10

–13.0

–20 –28.0

–30

Lower limit of agreement

–40 40.0

50.0

60.0 70.0 80.0 (MAP + DOP)/2 (mmHg)

90.0

100.0

Figure 2 Bland and Altman plot comparing Doppler measurements (DOP) with directly measured mean arterial pressure (DIR MAP) in 17 rabbits in which anaesthesia was maintained with isoflurane, totalling 190 comparisons. Each data point corresponds to the difference between a pair of Doppler and direct-measurements plotted against the mean of the two simultaneously obtained results. Horizontal lines indicate mean difference and the 95% limits of agreement. The trend line represents the change of mean difference across the measured range. A test of slope by regression analysis showed that this line was marginally significantly different from 0 (b = )0.108, SE = 0.054, t = )2.013, p = 0.046).

30

Figure 3 Graph showing the differences between directly measured systolic blood pressures (DIR SAP) and Doppler measurements (DOP) in 17 rabbits in which anaesthesia was maintained with isoflurane, in sequence of measurement during course of study, a total of consecutive 190 comparisons.

SAP – DOP (mmHg)

20 10 Mean difference +1.0

0 –10 –20 –30 –40 0

200 50 100 150 Sequence of measurement during course of study

rabbits appeared to require or were given extra analgesia during the recovery period. Discussion It is now widely accepted that the best way to assess the level of agreement between two methods of

250

measurement is to quantify the amount of disagreement between methods by calculating the differences between pairs of measurements. One of the methods of measurement is an accepted standard. The other method of measurement is a potentially useful alternative method being tested (Bland & Altman 2007). Differences between pairs

 2012 The Authors. Veterinary Anaesthesia and Analgesia  2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 174–184

179

Doppler blood pressure in anaesthetized rabbits L Harvey et al.

of measurements are calculated. The mean difference (often termed bias) is then calculated. Conventionally, differences are calculated by subtracting the measurements obtained from the test method from those of the accepted standard. A positive mean difference therefore indicates that the measurement method being evaluated tends to underestimate, whilst a negative mean difference indicates that the method being evaluated tends to overestimate. The standard deviation of the differences is also calculated (often termed precision) and this is used to calculate limits of agreement. A nonparametric form of the limits of agreement method is sometimes used, whereby differences are grouped into ranges, and the proportion of differences in each range is calculated (Bland & Altman 2007). Guidelines have been published by the Association for the Advancement of Medical Instrumentation (AAMI guidelines quoted by Stokes et al. 1991) and the European Hypertension Society (O’Brien et al. 2002) for the evaluation of blood pressure measurement devices in people. These set out the manner in which devices should be tested against the accepted standard as well as pass/fail criteria for devices. The AAMI guidelines recommend a mean difference of <5 mmHg with a standard deviation of <8 mmHg as an acceptable error when comparing non-invasive and direct blood pressure measurements. The level of agreement between Doppler measurements and directly measured SAP in this study was within this recommended minimum standard. The ESH groups comparisons between pairs of measurements (test device versus mercury sphygmomanometry with stethoscope auscultation of Korotkoff sounds) according to whether they are <5, 10 and 15 mmHg of each other. In the first phase of testing three comparisons are made from each of 15 subjects, giving 45 comparisons. For a device to pass, either at least 25 comparisons must lie within 5 mmHg, 35 must lie within 10 mmHg, or 40 must lie within 15 mmHg. In the second phase of testing, three comparisons are made from each of 33 subjects, giving 99 comparisons. For a device to pass, there must be a minimum of 60, 75 and 90 comparisons falling within 5, 10 and 15 mmHg, respectively. There must also be a minimum of either 65 comparisons within 5 mmHg and 80 comparisons within 10 mmHg, or 65 comparisons within 5 mmHg and 95 comparisons within 15 mmHg, or 80 comparisons within 10 mmHg and 95 comparisons within 15 mmHg. Comparing Doppler measurements and directly 180

measured SAP, the Doppler method in this study was extremely close to meeting these requirements of the EHS. Guidelines for validation of blood pressure measurement devices in dogs and cats have been recently published by the American College of Veterinary Internal Medicine (Brown et al. 2007). These ACVIM guidelines recommend taking multiple non-invasive blood pressure measurements and recording at least three consecutive consistent (<20% variability) values which are averaged to obtain a final measurement. The reference method may be direct arterial blood pressure measurement or another non-invasive blood pressure measurement technique. System efficacy is met if the mean difference is <10 mmHg with a standard deviation of <15 mmHg. Additionally, at least 50% of measurements must lie within 10% and 80% of measurements lie within 20 mmHg of the reference method and correlation should be >0.9. Comparing Doppler measurements and directly measured SAP, the Doppler method in this study is well within the ACVIM requirements, except for correlation. The above guidelines cannot all be applied precisely to this study. For example, the EHS guidelines have been designed for comparison of non-invasive blood pressure measurement devices against auscultatory sphygmomanometry and a larger number of subjects than used in this study are required. Furthermore, the EHS guidelines require triplicate data collection to allow assessment of repeatability in another phase of analysis. The ACVIM guidelines recommend collection of 5–7 consistent measurements which are then averaged to give a final measurement. This averaging makes it more likely that a blood pressure measurement technique will meet the criteria. Replication of measurements was not performed in this study as direct measurements were changing rapidly and did not allow time for multiple measurements to be completed without a change in a rabbit’s actual ABP. However, in cats, the first Doppler measurement of SAP is an excellent predictor of the mean of 5 readings (Jepson et al. 2005). We conclude that there was a good level of agreement between directly measured SAP and Doppler measurements in this study because the Doppler method met the level of agreement recommended in the AAMI guidelines, and was extremely close to meeting the requirements of both the EHS and the ACVIM. The level of agreement between directly measured MAP and Doppler measurements was poorer with

 2012 The Authors. Veterinary Anaesthesia and Analgesia  2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 174–184

Doppler blood pressure in anaesthetized rabbits L Harvey et al.

Doppler measurements tending to overestimate directly measured MAP. The findings of this investigation contrast with the results of other studies in different species. The Doppler technique has been reported to underestimate directly measured SAP in dogs and cats (Grandy et al. 1992; Binns et al. 1995; Caulkett et al. 1998; Haberman et al. 2006). Doppler measurements have been found to agree more closely with directly measured MAP than SAP in cats (Caulkett et al. 1998). Differences in study methods may in part explain these different findings, particularly differences in the site of direct ABP measurements, the cuff width: limb circumference ratio, positioning of the cuff and the use of headphones for detection of the Doppler flow signal to cut out varying levels of extraneous noise. Central ABP (such as in the aorta or carotid arteries) is higher than in the auricular artery in rabbits. An aorta–auricular SAP gradient of 11 mmHg in conscious rabbits (Wilson et al. 1975) and a carotid–auricular MAP gradient of 8 mmHg in rabbits anaesthetized with either halothane or isoflurane (Imai et al. 1999) have been reported. Ypsilantis et al. (2005) compared direct measurements of abdominal aortic and auricular blood pressures in anaesthetized rabbits. Correlation was high and the mean between-methods difference was small. Auricular ABP measurements increasingly underestimated abdominal aortic pressure as ABP increased. It is likely that the Doppler measurements recorded in the present study would slightly underestimate ABP at central sites. Most studies evaluating Doppler blood measurement in other species have compared directly measured central arterial pressures, e.g. from the femoral artery or terminal aorta, and Doppler measurement at peripheral sites (Binns et al. 1995; Caulkett et al. 1998). In the current study, however, both Doppler and direct measurements of ABP were both made peripherally. Auricular arterial pressures were chosen in this study as this site would be commonly used for direct ABP measurement clinically. Ideally the two techniques of ABP measurement should have been compared at the same sites in the arterial tree such that the true pressures at each site would be the same. The auricular artery and the dorsal carpal branch of the radial artery are, however, both peripheral vessels of similar size, in theory minimising differences. The ratio of cuff width: limb circumference ratio has been shown to affect oscillometric ABP

measurements in dogs (Sawyer et al. 1991). The rabbits in this study were small. Despite the use of the smallest commercially available cuffs, the cuff width: limb circumference ratio was higher in this study compared to other studies (Grandy et al. 1992; Binns et al. 1995; Caulkett et al. 1998; Haberman et al. 2006). This is likely to have affected our results. The level of agreement found in this study should not be expected if different cuff width: limb circumference ratios are used. In this study, the Doppler transducer was applied over the dorsal carpal branch of the radial artery. Studies in other species have found that Doppler measurements vary depending on the site used. Binns et al. (1995) reported that the Doppler technique on the hindlimb underestimated directly measured SAP in cats whereas applied to the tail it overestimated directly measured SAP. Agreement between Doppler and direct ABP measurements found in this current study cannot be extrapolated to other measurement sites. In several other studies, the level of agreement between indirect and direct blood pressure measurements changes over the measured range (Binns et al. 1995; Caulkett et al. 1998; Pederson et al. 2002). In this study, the mean betweenmethod difference was consistent across the measured pressure range when comparing direct SAP and Doppler measurements, i.e. there was no systematic change in the mean difference with pressure change. The mean between-method difference was found to change slightly across the measured pressure range when comparing direct MAP and Doppler measurements but this difference was very small (b = )0.108) and therefore unlikely to be of clinical significance and can be ignored. In the study by Caulkett et al. (1998), ABP was manipulated by adjusting the inspired concentration of isoflurane to produce a directly measured SAP range of approximately 40–100 mmHg. With the Doppler technique applied to the forelimb, Doppler underestimation of directly measured systolic blood pressure decreased with decreasing ABP. Binns et al. (1995) also manipulated the ABP in cats by altering anaesthetic depth and using vasopressors and positive inotropes, to give a wide ABP range. A ‘low pressure’ range was defined as a direct SAP <100 mmHg. With hindlimb Doppler measurements, Doppler underestimation of directly measured systolic blood pressure decreased with decreasing ABP. Conversely, the overestimation of directly measured systolic blood pressure by the

 2012 The Authors. Veterinary Anaesthesia and Analgesia  2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 174–184

181

Doppler blood pressure in anaesthetized rabbits L Harvey et al.

Doppler technique applied to the tail increased with decreasing pressure. Although the ABP range measured in the current study was limited, it is clinically relevant. Measures were taken to increase ABP if hypotension occurred, but as this was a clinical study, ABP was not manipulated experimentally. The ABP measured in the current study was low (directly measured SAP range 49–99 mmHg, directly measured MAP range 33–83 mmHg). The agreement between direct and Doppler measurements may differ outside of the studied ABP range. A limitation of this study is that the same person collected both DIR and DOP measurements and this person was not ‘blinded’ to the results during the course of the study. The subjective DOP measurement was taken before the DIR values were recorded from the monitor’s display and measurements were always made in this order. However, the person collecting this data could have become more skilled at interpreting the returning Doppler flow signal as the study progressed. The difference between pairs of measurements tended to decrease during the course of the study, showing that learning and improvement of technique occurred (see Fig. 3) despite considerable experience in using the Doppler method for arterial blood pressure measurement prior to study commencement. Consequently, the importance of experience should not be underestimated when using the Doppler technique for arterial blood pressure measurement. Interobserver differences in Doppler measurement determination are therefore a possibility, especially if there are large differences in observer experience. In cats, Jepson et al. (2005) found a small difference between the Doppler measured diastolic arterial blood pressure measurements but no difference between the Doppler measured systolic arterial blood pressure measurements made by two different observers. The level of experience of the investigators was not stated. Thermoregulation could cause changes in peripheral arteriolar vasomotor tone, possibly affecting the direct measurements of ABP measured at this site. In conscious rabbits, vasoconstriction of the middle and distal segments of the central ear artery has been shown to cause unpredictable temporal fluctuations in ABP measured directly in the central ear artery (Talseth et al. 1981). Isoflurane anaesthesia causes vasodilation and a reduced vasomotor response to changes in temperature. Measures were taken to prevent hypothermia and all the rabbits 182

had a rectal temperature >37.0 C at the end of anaesthesia. The administration of fluanisone is also likely to have reduced arteriolar vasomotor tone throughout anaesthesia. The application of EMLA cream may also have influenced arteriolar vasomotor tone (Bjerring et al. 1989). In all the rabbits, the auricular artery during anaesthesia was larger with a more palpable pulse when compared to during the pre-anaesthetic examination. Anaesthetic protocols using a-2 receptor agonist drugs (such as medetomidine) cause peripheral vasoconstriction. In this situation, the level of agreement between Doppler measured ABP in the forelimb and directly measured auricular ABP could be different to that found in this study. Excessive damping of the ABP waveform can be a problem with catheterisation of a small artery due to inadvertent kinking of the catheter, occlusion of the tip of the catheter against the wall of the artery, occlusion of the artery proximally due to positioning and the formation of clots within the catheter. This could have affected the accuracy of DIR ABP measurements against which the DOP measurements were compared; however, all data from a rabbit were excluded if excessive damping was not resolved. Calculation of the between-rabbit and withinrabbit variance showed that the majority of the total variability in the differences between methods in this study was due to between-rabbit factors. This may be caused by a less than perfect application of either technique but is most likely to be mainly due to small but unintentional differences in Doppler measurement technique. These will include factors beyond the control of the person taking a Doppler measurement, such as small variations in cuff width: limb circumference ratio owing to a lack of commercially available cuffs for patients of a range of sizes, audibility of flow signal, subtle differences in positioning of the limb in which the measurement is being taken owing to patient positioning for surgery, and variations in rabbit limb conformation. It is therefore obviously of great importance to take great care to avoid variations in technique that are avoidable, for example, selection of a different sized cuff. The use of headphones is recommended. Two different propofol preparations were used in this study due to unforeseen changes in drug purchasing. This should not have significantly affected the results of this study as each rabbit acted as its own control. Propofol is known to cause

 2012 The Authors. Veterinary Anaesthesia and Analgesia  2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 174–184

Doppler blood pressure in anaesthetized rabbits L Harvey et al.

pain on injection and to cause vasodilation, but any difference between the two products is unlikely to have affected auricular artery vasomotor tone. The oscillometric technique has been compared to direct measurements of ABP in the abdominal aorta in rabbits (Ypsilantis et al. 2005). Oscillometric measurements were made at two different sites, just proximal to the elbow and just proximal to the stifle and cuff width: limb circumference ratios were approximately 35% for forelimb and 30% for hindlimb measurements. Correlation coefficients and level of agreement varied, depending on the cuff position (forelimb or hindlimb) and ABP range. The best correlation was between oscillometric measurements with the cuff on the forelimb and directly measured SAP (r = 0.87) although oscillometric measurements underestimated direct measurements of ABP in the abdominal aorta. Correlation was poor using a cuff on the hindlimb. Unfortunately, the level of agreement between methods was displayed graphically but not provided numerically. Oscillometric techniques often fail to measure in very small patients (Kittleson & Olivier 1983; Binns et al. 1995; Caulkett et al. 1998), patients with high heart rates and severe hypotension (Ypsilantis et al. 2005) and in patients with small volume pulses or arrhythmias (Kittleson & Olivier 1983). Doppler measurements were obtained without difficulty in this study. The prevalence of hypotension was high in this study, in agreement with other studies of anaesthesia in rabbits (Imai et al. 1999; Martinez et al. 2009). To what extent this morbidity may contribute to the high anaesthetic-related mortality rate (1.39%) in rabbits (Brodbelt et al. 2008) is unknown, but it may be a factor. This study allows clinicians to use the Doppler technique for ABP measurement with knowledge of its accuracy and precision. In conclusion, the level of agreement between forelimb Doppler measured ABP and directly measured auricular SAP is good in rabbits during isoflurane anaesthesia. Doppler measurements below 80 mmHg are a highly reliable indicator of arterial hypotension as measured directly in the auricular artery. Acknowledgements The authors wish to thank our colleagues at the University of Bristol, particularly Alan Jones and Elisa Bortolami, for assistance with this study.

References American College of Veterinary Anesthesiology (1995) Suggestions for monitoring anesthetized veterinary patients. J Am Vet Med Assoc 206, 936–937. American Society of Anesthesiologists (2011) ASA Physical Status Classification. Available at: http:// www.asahq.org/clinical/physicalstatus.htm (Accessed 19 January 2011). Association of Anaesthetists of Great Britain and Ireland (2007) Recommendations for Standards of Monitoring During Anaesthesia and Recovery (4th edn). The Association of Anaesthetists of Great Britain and Ireland, London, UK. pp. 8–9. Bailey JE, Dunlop CI, Chapman PL et al. (1994) Indirect Doppler ultrasonic measurement of arterial bloodpressure results in a large measurement error in dorsally recumbent anesthetized horses. Equine Vet J 26, 70–73. Binns SH, Sisson DD, Buoscio DA et al. (1995) Doppler ultrasonographic, oscillometric sphygmomanometric, and photoplethysmographic techniques for noninvasive blood pressure measurement in anesthetized cats. J Vet Intern Med 9, 405–414. Bjerring P, Anderson PH, Arendt-Nielson L (1989) Vascular response of human skin after analgesia with EMLA cream. Br J Anaesth 63, 655–660. Bland JM, Altman DG (2007) Agreement between methods of measurement with multiple measurements per individual. J Biopharm Stat 17, 571–582. Branson KR (2007) Injectable and alternative anesthetic techniques. In: Lumb and Jones’ Veterinary Anesthesia and Analgesia (4th edn). Tranquilli WJ, Thurmon JC, Grimm KA (eds). Blackwell Publishing, IA, USA. pp. 273–300. Brodbelt DC, Blissitt KJ, Hammond RA et al. (2008) The risk of death: the confidential enquiry into perioperative small animal fatalities. Vet Anaesth Analg 35, 365– 373. Brown S, Atkins C, Bagley R et al. (2007) Guidelines for the identification, evaluation, and management of systemic hypertension in dogs and cats. J Vet Intern Med 21, 542–558. Caulkett NA, Cantwell SL, Houston DM (1998) A comparison of indirect blood pressure monitoring techniques in the anesthetized cat. Vet Surg 27, 370–377. Gaynor JS, Dunlop CI, Wagner AE et al. (1999) Complications and mortality associated with anesthesia in dogs and cats. J Am Anim Hosp Assoc 35, 13–17. Grandy JL, Dunlop CI, Hodgson DS et al. (1992) Evaluation of the Doppler ultrasonic method of measuring systolic arterial blood pressure in cats. Am J Vet Res 53, 1166–1169. Haberman CE, Kang CW, Morgan JD et al. (2006) Evaluation of oscillometric and Doppler ultrasonic methods of indirect blood pressure estimation in conscious dogs. Can J Vet Res 70, 211–217.

 2012 The Authors. Veterinary Anaesthesia and Analgesia  2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 174–184

183

Doppler blood pressure in anaesthetized rabbits L Harvey et al.

Herrold EM, Goldweit RS, Carter JN et al. (1992) Noninvasive laser-based blood pressure measurement in rabbits. Am J Hypertens 5, 197–202. Imai A, Steffey EP, Ilkiw JE et al. (1999) Comparison of clinical signs and hemodynamic variables used to monitor rabbits during halothane- and isofluraneinduced anesthesia. Am J Vet Res 60, 1189–1195. Jepson RE, Hartley V, Mendl M et al. (2005) A comparison of CAT Doppler and oscillometric Memoprint machines for non-invasive blood pressure measurement in conscious cats. J Feline Med Surg 7, 147–152. Kittleson MD, Olivier NB (1983) Measurement of systemic arterial blood-pressure. Vet Clin North Am Small Anim Pract 13, 321–336. Kurashina T, Sakamaki T, Yagi A et al. (1994) A new device for indirect blood pressure measurement in rabbits. Jpn Circ J 58, 264–268. Lemke KA (2007) Anticholinergics and sedatives. In: Lumb and Jones’ Veterinary Anesthesia and Analgesia (4th edn). Tranquilli WJ, Thurmon JC, Grimm KA (eds). Blackwell Publishing, Iowa, USA. pp. 203–240. Love L, Harvey R (2006) Arterial blood pressure measurement: Physiology, tools, and techniques. Contin Educ Anaesth Crit Care Pain 28, 450–461. Martinez MA, Murison PJ, Love E (2009) Induction of anaesthesia with either midazolam or propofol in rabbits premedicated with fentanyl/fluanisone. Vet Rec 164, 803–806. Morizono T, Johnstone BM, Lulich C (1975) Alternative method for measurement of arterial pressure in the rabbit. J Appl Physiol 39, 312–314. Newton GB, Mason RA, Hui JC et al. (1991) Indirect measurement of systolic blood pressure in the rabbit. J Invest Surg 4, 87–92. O’Brien E, Pickering T, Asmar R et al. (2002) Working Group on Blood Pressure Monitoring of the European Society of Hypertension International Protocol for validation of blood pressure measuring devices in adults. Blood Press Monit 7, 3–17.

184

Pederson KM, Butler MA, Ersboll AK et al. (2002) Evaluation of an oscillometric blood pressure monitor for use in anesthetized cats. J Am Vet Med Assoc 221, 646–650. Popesko P, Rajtova´ V, Hora´k J (2002) A colour atlas of anatomy of small laboratory animals. In: Volume 1: Rabbit - Guinea Pig. Saunders, Elsevier Science Ltd, Kidlington, Oxfordshire, UK. pp. 113. Sawyer DC, Brown M, Striler EL et al. (1991) Comparison of direct and indirect blood pressure measurement in anesthetized dogs. Lab Anim Sci 41, 134–138. Steffey EP, Mama KR (2007) Inhalation Anesthetics. In: Lumb and Jones’ Veterinary Anesthesia and Analgesia (4th edn). Tranquilli WJ, Thurmon JC, Grimm KA (eds). Blackwell Publishing, IA, USA. pp. 355– 394. Stokes DN, Clutton-Brock T, Patil C et al. (1991) Comparison of invasive and noninvasive measurement of arterial pressure using the Finapres. Br J Anaesth 67, 26–35. Talseth T, McNay JL, Ballinger K et al. (1981) Vasoconstriction limits to the use of the central ear artery pressure in conscious rabbits. Lab Anim 15, 1–3. Wagner AE, Brodbelt DC (1997) Arterial blood pressure monitoring in anesthetized animals. J Am Vet Med Assoc 210, 1279–1285. Ward M, Langton JA (2007) Blood pressure measurement. Contin Educ Anaesth Crit Care Pain 7, 122– 126. Wilson DM, Romero JC, Strong CG et al. (1975) Indirect blood pressure measurements in the rabbit: correlations with direct aortic and ear pressures. J Lab Clin Med 86, 1032–1039. Ypsilantis P, Didilis VN, Politou M et al. (2005) A comparative study of invasive and oscillometric methods of arterial blood pressure measurement in the anesthetized rabbit. Res Vet Sci 78, 269–275. Received 22 December 2010; accepted 23 May 2011.

 2012 The Authors. Veterinary Anaesthesia and Analgesia  2012 Association of Veterinary Anaesthetists and the American College of Veterinary Anesthesiologists, 39, 174–184