Clinical evaluation of the Surgivet V60046, a non invasive blood pressure monitor in anaesthetized dogs

Veterinary Anaesthesia and Analgesia, 2008, 35, 13–21

doi:10.1111/j.1467-2995.2007.00346.x

RESEARCH PAPER

Clinical evaluation of the Surgivet V60046, a non invasive blood pressure monitor in anaesthetized dogs Catherine JA Deflandre*

DVM, Cert VA

& Ludo J Hellebrekers 

DVM, PhD, Diplomate ECVA

*Department of Clinical Sciences, Small Animal Surgery, Faculty of Veterinary Medicine, University of Liege, Boulevard de Colonster, Liege, Belgium  Department of Clinical Sciences of Companion Animals, Faculty of Veterinary Medicine, University of Utrecht, Yalelaan, Utrecht, The Netherlands

Correspondence: Catherine JA Deflandre, De´partement des Sciences Cliniques, CHIPA, B44, Boulevard de Colonster, 20, Liege, 4000, Belgium. E-mail: [email protected]

Abstract Objective To compare the performance of the Surgivet Non-Invasive Blood Pressure (NIBP) monitor V60046 with an invasive blood pressure (IBP) technique in anaesthetized dogs. Study design A prospective study. Animals Thirty-four dogs, anaesthetized for a variety of procedures. Methods Various anaesthetic protocols were used. Invasive blood pressure measurement was made using a catheter in the femoral or the pedal artery. A cuff was placed on the contralateral limb to allow non invasive measurements. Recordings of arterial blood pressures (ABPs) were taken at simultaneous times for a range of pressures. For analysis, three pressure levels were determined: high [systolic blood pressure (SAP) > 121 mmHg], normal (91 mmHg < SAP < 120 mmHg) and low (SAP < 90 mmHg). Comparisons between invasive and non invasive measurements were made using Bland-Altmann analysis. Results The NIBP monitor consistently underestimated blood pressure at all levels. The lowest biases and greatest precision were obtained at low and normal pressure levels for SAP and mean arterial pressure (MAP). At low blood pressure levels, the biases ± 95%

confidence interval (CI) were 1.9 ± 2.96 mmHg (SAP), 8.3 ± 2.41 mmHg diastolic arterial pressure (DAP) and 3.5 ± 2.09 mmHg (MAP). At normal blood pressure levels, biases and CI were: 1.2 ± 2.13 mmHg (SAP), 5.2 ± 2.32 mmHg (DAP) and 2.1 ± 1.54 mmHg (MAP). At high blood pressure levels, the biases and CI were 22.7 ± 5.85 mmHg (SAP), 5.5 ± 3.13 mmHg (DAP) and 9.4 ± 3.52 mmHg (MAP). In 90.6% of cases of hypotension (MAP < 70 mmHg), the low blood pressure was correctly diagnosed by the Surgivet. Conclusions Measurement of blood pressure with the indirect monitor allowed detection of hypotension using either SAP or MAP. The most accurate readings were determined for MAP at hypotensive and normal levels. The monitor lacked accuracy at high pressures. Clinical relevance When severe challenges to the cardiovascular system are anticipated, an invasive method of recording ABP is preferable. For routine usage, the Surgivet monitor provided a reliable and safe method of NIBP monitoring in dogs, thereby contributing to the safety of anaesthesia by providing accurate information about the circulation. Keywords anaesthesia, arterial blood pressure, dog, hypotension, non invasive blood pressure, oscillometric.

13

Comparison of IBP and a NIBP method in dogs CJA Deflandre and LJ Hellebrekers

Introduction Arterial blood pressure (ABP) is a fundamental cardiovascular measurement that describes the force driving tissue perfusion. It is the most important determinant of left ventricular afterload and therefore, of the workload of the heart (Mark & Slaughter 2005) and one of the most commonly utilized indices of cardiovascular performance. At the tissue level, the fluctuations in arterial pressure are damped, so it is the mean arterial pressure (MAP) that determines how well tissues are perfused (Young 1999). The usefulness of ABP monitoring during anaesthesia has been known since 1897 when Hill and Barnard demonstrated a fall in blood pressure during chloroform anaesthesia (Navqvi 1998). Hypotension, defined as a mean blood pressure of <70 mmHg, is one of the most common anaesthetic complications both in human and veterinary anaesthesia (Paddleford 1992; Cockings et al. 1993; Gaynor et al. 1999; Wagner et al. 2003). Unfortunately, ABP is not commonly monitored in veterinary practice (Dyson et al. 1998; Nicholson & Watson 2001; Wagner & Hellyer 2002). In human anaesthesia, monitoring of ABP is part of the basic standard of monitoring of the American Society of Anaesthesia and of the Association of Anaesthetists of Great Britain and Ireland (Association of Anaesthetists of Great Britain and Ireland 2000; American Society of Anesthesiologists 2004). The American College of Veterinary Anesthesiology published similar guidelines in 1995 in which ABP monitoring was only recommended for feline and canine patients assessed to be ASA status III-V (American College of Veterinary Anesthesiology 1995) or in equine anaesthesia when inhalational agents were used (Martinez et al. 2005). There are many ways to assess the function of the cardiovascular system, but not all of them are useful indicators of tissue perfusion or MAP. Palpation of the peripheral pulse, evaluation of the fade in the Doppler’s audio output and alterations in heart rate were found to be unreliable predictors of hypotension (Waelchli-Suter et al. 1986; Paddleford 1992; Dyson 1997; Wagner & Brodbelt 1997). Of the two methods of ABP measurement, the direct or invasive method (IBP), is the accepted standard and the most accurate, especially in hypo- and hypertensive states (Wagner & Brodbelt 1997) and in addition, it provides continuous measurement. However, the placement of an arterial catheter in 14

small patients can be challenging. This technique is also associated with slight risks, such as haemorrhage, infection, partial or complete occlusion with distal ischemia, arterial embolization and an increase in anaesthetic time. Therefore, a reliable method of estimating ABP that does not involve arterial catheterization would be valuable. Two main non invasive ABP monitoring techniques currently used in veterinary medicine are the Doppler ultrasonic flow detector and oscillometric sphygmomanometry. Numerous comparative studies of the techniques are available, but variable results have been reported. In cats (Klevans et al. 1979; Binns et al. 1995; Caulkett et al. 1998) and also in dogs (Garner et al. 1975), the Doppler was found to have a high degree of accuracy. In cats, an adjustment factor of 14 mmHg should be added to the Doppler systolic pressure to obtain the direct systolic blood pressure (SAP) (Grandy et al. 1992). A good correlation was found with IBP readings for the oscillometric methods in dogs, with the cuff placed on the tail (Bodey et al. 1994), or on the pelvic limb (Weiser et al. 1977; Hamlin et al. 1982; Gains et al. 1995; Sawyer et al. 2004). Geddes et al. (1980) reported a good agreement with IBP for MAP when cuff width was 40% of the thoracic limb circumference. The 40% value is now commonly recommended when choosing an adequate cuff size for dogs and a 30% value for cats (Sawyer et al. 1991; Grandy et al. 1992). Grosenbaugh & Muir (1998) tested two oscillometric Non-Invasive Blood Pressure (NIBP) monitors and found that both were generally more accurate at normal and low ABP, but both had a tendency to underestimate direct readings, as did the monitor tested by Sawyer et al. (2004). Of the two monitors evaluated by Hunter et al. (1990), only one was of sufficient accuracy to be useful for measurement of NIBP in dogs (1990). The unsuitable monitor in that study was not calibrated for use with infant cuffs, highlighting again the need for a correct cuff size. In conscious dogs, Stepien & Rapoport (1999) reported poor agreement between IBP and both oscillometry and Doppler ultrasonography, with MAP, the closest to its directly measured counterpart. Another study in conscious dogs found that placing the cuff on the tail with the dog in standing position provided sufficient accuracy with minimal stress to the patient (Bodey et al. 1996). Single measurement must be interpreted cautiously and repeated measurements over a short period of time have been shown to increase correlation between IBP and

 2007 The Authors. Journal compilation  2007 Association of Veterinary Anaesthetists, 35, 13–21

Comparison of IBP and a NIBP method in dogs CJA Deflandre and LJ Hellebrekers

NIBP (Hunter et al. 1990; Bodey et al. 1996; Meurs et al. 1996). The Surgivet V60046 is a compact, portable battery or mains-operated monitor. It uses oscillometric principles in which an inflated cuff, placed around the patient’s limb, inflates and deflates in cycles to calculate the systolic, diastolic and mean ABP values. The cuffs used with this monitor are Purple Cuff (Technicuff, Leesburg, FL, USA). With these cuffs, sound as well as pressure is used to obtain readings. The inflating bladder covers the whole length of the cuff, therefore eliminating the need for alignment of the artery with the inflating tube. Each cuff also covers a wider range of limb size, minimizing the number of different cuffs necessary. This study was conducted to evaluate the accuracy of the Surgivet V60046 Non-Invasive Blood Pressure monitor (SurgiVet, Inc, Waukesha, WI, USA) in dogs under general anaesthesia. Materials and methods Thirty-four dogs, of various breeds, weighing between 3.7 and 52 kg (25.0 kg, 15.0 kg, 2.6 kg; mean, SD, SEM, respectively) were anaesthetized with commonly used anaesthetic protocols, chosen according to the physical status of the dog. The clinical condition of the dogs warranted invasive blood pressure (IBP) monitoring and in none of them was the artery catheterized solely for the purpose of the study. Ten dogs received no premedication, twelve received methadone (Methadone HCl, Eurovet Animal Health, Bladel, The Netherlands) and midazolam (Dormicum, Roche Nederland BV, Woerden, The Netherlands), seven received methadone only, one received midazolam and one received acepromazine and methadone. Anaesthesia was induced with propofol (Propovet, Abbott Animal Health, Madrid, Spain) (n ¼ 19), midazolam and sufentanil (n ¼ 10), thiopental (Nesdonal, Merial BV, Amstelveen, The Netherlands) or ketamine (Narketan, Vetoquinol BV, ‘s-Hertogenbosch, The Netherlands) (n ¼ 1, each). Anaesthesia was maintained with a combination of isoflurane (Isoflo, Abbott Animal Health, UK) delivered in oxygen and air. An additional continuous rate infusion of fentanyl (Fentanyl-B, Braun, Melsungen, Germany) (n ¼ 1), sufentanil (Sufentanil-hameln, Hameln pharmaceuticals, Hameln, Germany) (n ¼ 16) or sufentanil and midazolam (n ¼ 11) was administered in specific cases. Dogs were connected to an Aestiva 5 anaesthetic

machine (Datex-Ohmeda Division Instrumentarium Corp, Helsinki, Finland) using an appropriate breathing system (Ayres T-Piece or Circle). The electrocardiogram, haemoglobin saturation (SpO2), plethysmograph, spirometer, end-tidal (Et) isoflurane, end-tidal carbon dioxide (PE¢CO2), inspired fraction (Fi) of O2 and Fi isoflurane were monitored using a Datex-Ohmeda S/5 monitor (Datex-Ohmeda Division, Instrumentarium Corp, Helsinki, Finland). All of the animals received fluids administered intravenously, mainly lactated ringers (B Braun, Melsungen, Germany), but also Haes-steril 6% (Fresenius Kabi, ‘s-Hertogenbosch, The Netherlands) or plasma where indicated. Routine monitoring also included urine production and assessment of depth of anaesthesia. When necessary, mechanical ventilation was applied to maintain PE¢CO2 between the values of 4.5 and 6.0 kPa. Duration of anaesthesia was between 50 and 289 minutes (168.1, 65.99, 11.15; mean, SD, SEM). The direct or IBP was measured using a catheter placed aseptically in either the dorsal pedal or the femoral artery (Abbocath-T, 22 SWG, 32 mm, Abbot, Sligo, Republic of Ireland, Arterial cannula with flowswitch, 20 SWG, 45 mm, Braun Dickinson, Swindon, UK or Secalon T over the needle central venous catheter with flowswitch, 18 SWG, 90 mm, Braun Dickinson Critical Care Systems Pte Ltd, Singapore). The catheter was connected to a disposable transducer (Gabarith TM single transducer set, Braun Dickinson Infusion Therapy Systems Inc, Sandy, UT, USA) either directly or using an extension (Combydin Pressure Tubing, B Braun, Metsungen, Germany). The transducer was positioned at either the level of the point of the shoulder with dogs in dorsal recumbency and at mid-sternal level with dogs in lateral recumbency and zeroed at that level. The catheter was flushed regularly using NaCl 0.9% (B Braun, Melsungen, Germany) with heparin (5 IU ml)1). The transducer was connected to a Datex-Ohmeda S/5 monitor. After measurement of the circumference of the limb, an appropriate cuff size was chosen according to the manufacturer recommendation. Cuffs were placed either on the contra-lateral leg, just distal to the hock, or in two cases on the metacarpus. Three cuff sizes were available: small (recommended leg circumference 3–9 cm), medium (recommended leg circumference 5–15 cm) and large (recommended leg circumference 9–25 cm). Hair was not clipped over the measurement site.

 2007 The Authors. Journal compilation  2007 Association of Veterinary Anaesthetists, 35, 13–21

15

Comparison of IBP and a NIBP method in dogs CJA Deflandre and LJ Hellebrekers

Non-invasive blood pressure was measured at intervals of no less than 5 minutes, but simultaneous recording was performed occasionally every 10 minutes, when patient care prohibited more frequent readings. Simultaneous recording of the Surgivet data and the IBP data was made by hand. Care was taken to ensure that the direct measurements coincided with the NIBP measurement cycle. Data were then transferred into an Excel-spread sheet. Readings were organized into three groups: high SAP (above 121 mmHg), normal SAP (between 91 and 120 mmHg) and low SAP (lower than 90 mmHg). In each group, dogs with at least four measurements for the specific ABP level were selected. When more than four readings were available, four were chosen at random by an independent observer. Mean, standard deviation (SD) and standard error of the mean (SEM) were calculated for each pressure level. The agreement of the non invasive SAP, diastolic arterial pressure (DAP) and MAP with the corresponding simultaneous invasive ABP values was analyzed according to the Bland-Altman method (Bland & Altman 1986,1995,2003). The mean difference (bias) and SD of the difference (precision) were calculated by subtracting the values from the Surgivet technique from the corresponding direct arterial pressure values. A positive bias reflected an underestimation by the Surgivet monitor and a negative bias, an overestimation of the direct blood pressure value. The normal distribution of the values was checked by drawing a histogram. The limits of agreement for the bias estimates the range of agreement between the two techniques with a 95% confidence interval (CI) (2 SD). Bland-Altman plots of the differences in measurements (direct–indirect) against the mean of each pair of readings were drawn. Lines representing each bias and two SDs, above and below the mean, are provided. These boundaries have been suggested as acceptable limits of the departure from the directly measured values (Bland & Altman 1986). For each pressure group, linear regression was performed and regression equations established. In this model, the direct pressures were the dependent variables and the indirect pressures were the independent variables. These equations provide a mean of predicting direct pressure values from the readings of the Surgivet monitor. The MAP was chosen for inspection of the tracking capability of the NIBP monitor. All 473 readings were considered. A change in IBP larger than or equal to 16

10 mmHg was observed on 125 occasions and the corresponding change in NIBP was evaluated. Values from both methods should change overall in the same direction. From this were determined how close the indirect were associated with the direct measurements. Statistical analysis was performed using Excel with Analyze-it and SPSS.12.0.1. Results A variable number of sequential measurements were made in each dog, depending on the duration of surgery (min 6 – max 30, 15.2, 5.80, 1.04; mean, SD, SEM, respectively). In total, 473 readings were recorded. The grouping of readings according to blood pressure values was 13 dogs or 52 readings in the low ABP group, 24 dogs or 96 readings in the normal ABP group and 9 dogs or 36 readings in the high ABP group. The oscillometric monitor failed to provide a measurement in only 0.006% of readings. The lowest MAP read by the Surgivet was 42 mmHg, which corresponded to a direct MAP of 40 mmHg. At each failure, to obtain a reading, the Surgivet gave a measurement at the next attempt. The Positive Predictive value for the monitor was 90.6%, or the monitor correctly predicted hypotension in 90.6% of the cases. The sensitivity, which was defined as the correct identification of a direct MAP < 70 mmHg was 82.8% and the specificity, defined as the correct identification of a direct MAP > 70 mmHg was 91.2%. For each group of pressures, the comparative data are presented in Tables 1, 2 and 3. The Surgivet underestimated ABP at all levels. The most notable variations following ABP changes were seen for SAP; for which bias and precision increased from hypotension to hypertension and for which ABP determinations were most accurate during hypotension. The best agreement between the two methods was found for MAP. Mean arterial pressure was not very dependent on ABP level, but nevertheless showed an increased bias and precision at high ABP. The Surgivet constantly underestimated direct MAP between 2 and 5.5 mmHg. Additionally, at normal and low pressures, the SDs were less than 8 mmHg. The limits of agreements for MAP were +19.5 and )13.1. Bias and precision for DAP were intermediate to those calculated for MAP and SAP. Diastolic arterial pressure measurements appeared least dependent on changes in blood pressure and

 2007 The Authors. Journal compilation  2007 Association of Veterinary Anaesthetists, 35, 13–21

Comparison of IBP and a NIBP method in dogs CJA Deflandre and LJ Hellebrekers

Table 1 Systolic arterial blood pressure presented by pressure group

Pressure group*

n 

Direct, Mean ± SEM

Indirect, Mean ± SEM

Bias

Precision

95% CI

Correlation coefficient

High Normal Low Overall

36 96 52 184

141 104 78 104

118 95 77 95

22.7 8.6 1.9 9.2

17.90 11.89 11.06 14.50

5.85 2.13 2.96 2.09

0.49 0.39 0.65 0.80

± ± ± ±

2.7 0.9 1.1 1.8

± ± ± ±

2.9 1.2 1.9 1.5

CI, Confidence interval. *High: SAP > 121 mmHg; normal: 91 mmHg < SAP < 120 mmHg; low: SAP < 90 mmHg;  number of readings, from nine dogs in high BP group, 24 dogs in normal BP group and 13 dogs in low BP group.

Table 2 Diastolic blood pressure presented by pressure group

Pressure group*

n 

Direct, Mean ± SEM

Indirect, Mean ± SEM

Bias

Precision

95% CI

Correlation coefficient

High Normal Low Overall

36 96 52 184

76 59 44 58

67 53 35 51

9.4 5.2 8.2 6.9

10.78 12.10 8.85 11.10

3.52 1.57 2.09 1.60

0.76 0.66 0.50 0.80

± ± ± ±

2.0 1.1 0.9 1.1

± ± ± ±

2.6 1.6 1.4 1.4

CI, Confidence interval. *High: SAP > 121 mmHg; normal: 91 mmHg < SAP<120 mmHg; low: SAP < 90 mmHg;  number of readings, from nine dogs in high BP group, 24 dogs in normal BP group and 13 dogs in low BP group.

Table 3 Mean blood pressure presented by pressure group

Pressure group*

n 

Direct, Mean ± SEM

Indirect, Mean ± SEM

Bias

Precision

95% CI

Correlation coefficient

High Normal Low Overall

36 96 52 184

94 71 56 71

88 69 53 68

5.5 2.1 3.4 3.1

9.58 0.79 7.68 8.32

3.13 2.32 2.41 1.20

0.79 0.74 0.65 0.87

± ± ± ±

2.1 1.0 0.9 1.2

± ± ± ±

2.5 1.1 1.4 1.3

CI, Confidence interval. *High: SAP > 121 mmHg; normal: 91 mmHg < SAP<120 mmHg; low: SAP < 90 mmHg;  number of readings, from nine dogs in high BP group, 24 dogs in normal BP group and 13 dogs in low BP group.

pulse quality even if the precision was greatest during hypotension. Bland-Altman plots are shown in Fig. 1a–c, with lines at the mean bias and two SDs, above and below the mean. They confirm the increase in deviation from the invasive measurements seen at high ABP for SAP. The range between the upper and lower limits was 56 mmHg for SAP, 44 mmHg for DAP and 32 mmHg for MAP. The scatter of points above and below the line widens with the increase in ABP for SAP and MAP, demonstrating that the accuracy decreases with high pressures. Fig. 2a–c represents scatter plots with fitted regression lines, CI band and prediction interval band for a 95% CI. Systolic arterial pressure and MAP regression slopes were the closest to unity

(R ¼ 1), with R ¼ 0.946 and 0.876, respectively. Using the regression equations, the corresponding direct value can be calculated from the measurement given by the Surgivet using the following formulae: SAP direct ¼ (0.946 · Indirect SAP) + 14.36, MAP direct ¼ (0.827 · Indirect MAP) + 14.98 and DAP direct ¼ (0.656 · Indirect DAP) + 24.55. In 108 of 125 changes (86%) in ABP larger than or equal to 10 mmHg, the direction of change of NIBP agreed with the direction of change of IBP. In five of these tracking errors, even if the change was in the wrong direction, it brought, in fact, the NIBP reading closer to the IBP value. Table 4 shows the numbers and percentages of readings organized by discrepancy levels; 10, 15 and 20 mmHg. Respectively,

 2007 The Authors. Journal compilation  2007 Association of Veterinary Anaesthetists, 35, 13–21

17

Comparison of IBP and a NIBP method in dogs CJA Deflandre and LJ Hellebrekers

Direct systolic arterial blood pressure (mmHg)

(a)

30 20 10 0 –10 –20 –30 –40 –50 –60 20

Zero bias

40

60

80 100 120 140 160 180 200 Mean SAP (mmHg)

30 20

Y=X 180

y = 0.9462x + 14.361

160 140 120 100 80 60 40

20

Zero bias

0 –10 –20

(b)

–30 –40 –50 –60 20

40

60

80

100 120 140 160 180 200

Mean DAP (mmHg)

(c)

Difference in MAP between IBP and NIBP (mmHg)

Line of Identity

200

20

10

Direct diastolic arterial blood pressure (mmHg)

Difference in DAP between IBP and NIBP (mmHg)

(b)

30 20 10 0 –10 –20 –30 –40 –50 –60

Zero bias

20

40

60

80

100

120

140

160

180

Mean MAP (mmHg)

Figure 1 Bland-Altman plots of agreement between the Surgivet and the IBP. Plot of the difference against the mean of the differences (Direct-Surgivet). Dashed lines indicate mean difference ± 2 SD, dotted lines indicate mean bias. (a) SAP, (b) DAP, (c) MAP.

56%, 77% and 78% of SAP, DAP and MAP indirect readings were within 15 mmHg of the direct readings.

40 60 80 100 120 140 160 180 Indirect systolic arterial blood pressure (mmHg)

200

Line of Identity Y=X

120 y = 0.656x + 24.552

100

80

60

40

20 0 0 60 20 40 80 100 120 Indirect diastolic arterial blood pressure (mmHg)

200

(c)

Line of Identity

140

Y=X Direct mean arterial blood pressure (mmHg)

Difference in SAP between IBP and NIBP (mmHg)

(a)

y = 0.8269x + 14.98

120

100

80

60

40

Discussion 20

As ABP measurement is used in anaesthesia to make important therapeutic decisions regarding fluid therapy and use of inotropic or vasopressor agents, it is crucial that methods used yield reliable results. Standards for performance of automated NIBP devices have been set by the American Association of Medical instrumentation (AAMI) and the British Hypertension Society. These standards require NIBP monitors to yield measurements within 5 ± 8 mmHg (mean ± SD) of prediction 18

20

40

60

80

100

120

120

Indirect mean arterial blood pressure

Figure 2 Scatter plots of non invasive pressures versus direct pressures for overall measures. (a) Systolic arterial pressure with best-fit regression line, (b) Diastolic arterial pressure with best-fit regression line and (c) Mean arterial pressure with best-fit regression line. Outside lines represent the prediction interval band (95% CI), the dotted lines represent the confidence interval band (95% CI) and the thick line the fitted regression line.

 2007 The Authors. Journal compilation  2007 Association of Veterinary Anaesthetists, 35, 13–21

Comparison of IBP and a NIBP method in dogs CJA Deflandre and LJ Hellebrekers

Table 4 Overall frequency (number and percentage) of different discrepancy ranges between oscillometric and intra-arterial blood pressure measurement

Pressure variable

Discrepancy < 10 mmHg, n (%)

Discrepancy < 15 mmHg, n (%)

Discrepancy < 20 mmHg, n (%)

SAP DAP MAP

175 (36) 260 (53) 332 (68)

274 (56) 372 (77) 377 (78)

332 (68) 386 (80) 434 (89)

SAP, systolic blood pressure; DAP, diastolic arterial pressure; MAP, mean arterial pressure.

error. Few veterinary studies of NIBP monitors have met this standard (Binns et al. 1995). The Surgivet monitor reached the AAMI standard for MAP at low and normal ABP and was very close to this standard for the overall MAP. Pulse detection is one of the most important factors affecting accuracy of NIBP devices (Grosenbaugh & Muir 1998). Mean arterial pressure being the largest signal received, it is the most accurate value measured using the oscillometric method. In addition, using direct ABP measurement methods, MAP is less likely to be altered by distortions generated by recording systems and dynamic response characteristics of the systems are of little importance (Gardner 1981). The increased difference between IBP and NIBP measurements at high pressure states, especially for SAP (Henneman & Henneman 1989), can be accounted for by various phenomena. In hypertensive states, the peak of the pulse wave is sustained for a short time only and may not be detected by flow-dependent sensors in the cuff, thus falsely recording a lowering of SAP (Binns et al. 1995; Grosenbaugh & Muir 1998). The tendency for the NIBP method to underestimate can be also, in part, attributed to the substantial amount of energy required to generate oscillations in cuff pressure. This is contrary to the situation with pressure transducers which can record a peak pressure without a great loss in energy (Binns et al. 1995). In addition, apparent underestimation by the NIBP method could be as a result of the direct ABP being artificially increased via vasoconstriction and increased peripheral resistance and via an increased mismatch of patient and recording system impedance (Grosenbaugh & Muir 1998). If the resonant frequency is too close to the frequency of the signal

being measured, then distortion of the high frequency part of the pressure pulse will occur, causing an increased artefact in systolic pressure measurement and the increased variance between IBP and NIBP signal at SAP and high ABP (Bruner et al. 1981a,b; Kleinman 1989). In our study, accuracy was indeed decreased at hypertension and for SAP. The tendency of the Surgivet machine to underestimate ABP is in general agreement with other reports on NIBP (Geddes et al. 1980; Bodey et al. 1994; Binns et al. 1995; Grosenbaugh & Muir 1998; Sawyer et al. 2004). The main issue when comparing direct and indirect measurements is that two inherently different phenomena are being compared based on the assumption that blood flow and ABP are directly related (Henneman & Henneman 1989). In effect, IBP measurement is pressure-dependent and the oscillometric method is blood flow-dependent. Inferences about flow based on pressure are only valid if resistance remains constant. However, peripheral resistance varies with the tissue’s perfusion needs, and BP and blood flow are not always well correlated. Another major issue when comparing the two methods is the assumption that the IBP is the absolute standard, therefore giving ‘the true ABP’. Although precautions were taken, IBP is also subject to inaccuracies. Primarily, the relationship between damping of the pulse wave and the frequency response of the monitoring system will determine how accurately the pressure pulse is transmitted. Damping will lower this frequency response (Henneman & Henneman 1989). We did not calibrate our pressure transducers against a known standard (mercury manometer) so the accuracy of the transducers was not verified for each animal adding a further unknown discrepancy to our results. It is also unfortunate that the measurements were made at different sites by each method, as it cannot be excluded that differences in ABP existed between opposite limbs. It would, obviously, have been better to take both direct and indirect measurements from the same site, but this was not feasible. Some studies have found the pelvic limbs to provide better accuracy (Bodey et al. 1994), but it is not a constant feature (Branson et al. 1997). We did not attempt to evaluate the interaction between cuff placement and accuracy of measurements. Comparisons of these results with those of previous studies are hindered by the variety of combinations of artery and transducer system used and cuff application sites. Some studies also used a

 2007 The Authors. Journal compilation  2007 Association of Veterinary Anaesthetists, 35, 13–21

19

Comparison of IBP and a NIBP method in dogs CJA Deflandre and LJ Hellebrekers

smaller range of patient sizes (Gains et al. 1995; Grosenbaugh & Muir 1998; Sawyer et al. 2004). Precision was improved in other studies by taking serial measurements and averaging the readings (Bodey et al. 1996; Meurs et al. 1996). This is partly as a result of the possible large short-term variations of the SAP and DAP. Cardiovascular depression is to a variable extent always a feature of general anaesthesia as a result of anaesthetic drugs, vasodilatation, blood loss or vascular compromise. As MAP, rather than SAP or DAP, is the true driving pressure for tissue perfusion, it is recommended that clinical decisions be made and that therapy should be based on MAP (Tibby 2002). Prolonged hypotension, below a MAP of 60 mmHg, even when of a moderate degree, may lead to compromised renal, cardiac and cerebral perfusion. Measurement of blood pressure by a NIBP method, if proved reliable, would allow early detection of any hypotensive episode, its prompt treatment and evaluation of its effectiveness. There are no statistical data regarding the benefit of sophisticated intraoperative monitors on the incidence of adverse events in veterinary anaesthesia. A parallel can nevertheless be drawn with human anaesthesia, and with the Australian Incident Monitoring Study in which the authors suggest a priority list for monitor acquisition for those with limited resources, placing NIBP in second place after the stethoscope (Webb et al. 1993). Thus, from a clinical perspective and by standards used in human anaesthesia, the Surgivet with a cuff placed on the metatarsal region would be acceptable for measurement of MAP in dogs. When using MAP as a reference for detection of normo- or hypotension in anaesthetized patients, the Surgivet monitor produced values that were within close range of those obtained directly. The advantages of this type of monitor compensate for the loss of accuracy, as the inaccuracies are predictable and the trends observable. This monitor can therefore be used to increase the safety of anaesthesia. In animals, in which immediate and reliable detection of changes in ABP is critical, invasive monitoring should be used. The increased magnitude of error at high pressures may also be important when using the device for the monitoring of hypertensive states. The regression coefficient close to unity obtained for SAP means that, for ease of clinical use, the additive value of 14 mmHg can be added to estimate the direct SAP from the indirect value. 20

Acknowledgements This work was made possible by the donation of the NIBP monitor by Smith Medical PM, Inc. of Waukesha, WI, USA via Guido J.M.M. de Nijs of Denijs Educatie, Asch, The Netherlands. 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 (2004) Standards for basic anesthetic monitoring. http://www.asahq.org/ publicationsAndServices/standards/02.pdf. Association of Anaesthetists of Great Britain and Ireland (2000) Recommendations for standards of monitoring during anaesthesia and recovery. http://www.aagbi. org/pdf/Absolute.pdf. 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. Bland JM, Altman DG (1986) Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1, 307–310. Bland JM, Altman DG (1995) Comparing methods of measurement: why plotting difference against standard method is misleading. Lancet 346, 1085–1087. Bland JM, Altman DG (2003) Applying the right statistics: analyses of measurement studies. Ultrasound Obstet Gynecol 22, 85–93. Bodey AR, Young LE, Bartram DH et al. (1994) A comparison of direct and indirect (oscillometric) measurements of arterial blood pressure in anaesthetised dogs, using tail and limb cuffs. Res Vet Sci 57, 265–269. Bodey AR, Michell AR, Bovee KC et al. (1996) Comparison of direct and indirect (oscillometric) measurements of arterial blood pressure in conscious dogs. Res Vet Sci 61, 17–21. Branson KR, Wagner-Mann CC, Mann FA (1997) Evaluation of an oscillometric blood pressure monitor on anesthetized cats and the effect of cuff placement and fur on accuracy. Vet Surg 26, 347–353. Bruner JMR, Krenis LJ, Kunsman JM et al. (1981a) Comparison of direct and indirect methods of measuring arterial blood pressure. Med Instrum 15, 11–21. Bruner JMR, Krenis LJ, Kunsman JM et al. (1981b) Comparison of direct and indirect methods of measuring arterial blood pressure, part II. Med Instrum 15, 97–101. Caulkett NA, Cantwell SL, Houston DM (1998) A comparison of indirect blood pressure monitoring techniques in the anesthetized cat. Vet Surg 27, 370–377.

 2007 The Authors. Journal compilation  2007 Association of Veterinary Anaesthetists, 35, 13–21

Comparison of IBP and a NIBP method in dogs CJA Deflandre and LJ Hellebrekers

Cockings JG, Webb RK, Klepper ID et al. (1993) The Australian Incident Monitoring Study. Blood pressure monitoring – applications and limitations: an analysis of 2000 incident reports. Anaesth Intensive Care 21, 565– 569. Dyson DH (1997) Assessment of 3 audible monitors during hypotension in anesthetized dogs. Can Vet J 38, 564– 566. Dyson DH, Maxie MG, Schnurr D (1998) Morbidity and mortality associated with anesthetic management in small animal veterinary practice in Ontario. J Am Anim Hosp Assoc 34, 325–335. Gains MJ, Grodecki KM, Jacobs RM et al. (1995) Comparison of direct and indirect blood pressure measurements in anesthetized dogs. Can J Vet Res 59, 238–240. Gardner RM (1981) Direct blood pressure measurementdynamic response requirements. Anesthesiology 54, 227–236. Garner HE, Hahn AW, Hartley JW et al. (1975) Indirect blood pressure measurement in the dog. Lab Anim Sci 25, 197–202. 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. Geddes LA, Combs W, Denton W et al. (1980) Indirect mean arterial pressure in the anesthetized dog. Am J Physiol 238, H664–H666. 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. Grosenbaugh DA, Muir WW (1998) Accuracy of noninvasive oxyhemoglobin saturation, end-tidal carbon dioxide concentration, and blood pressure monitoring during experimentally induced hypoxemia, hypotension, or hypertension in anesthetized dogs. Am J Vet Res 59, 205–212. Hamlin RL, Kittleson MD, Rice D et al. (1982) Noninvasive measurement of systemic arterial pressure in dogs by automatic sphygmomanometry. Am J Vet Res 43, 1271–1273. Henneman EA, Henneman PL (1989) Intricacies of blood pressure measurement: Reexamining the rituals. Heart Lung 18, 263–273. Hunter JS , McGrath CJ, Thatcher CD et al. (1990) Adaptation of human oscillometric blood pressure monitors for use in dogs. Am J Vet Res 51, 1439–1442. Kleinman B (1989) Understanding natural frequency and damping and how they relate to the measurement of blood pressure. J Clin Monit 5, 137–147. Klevans LR, Hirkaler G, Kovacs JL (1979) Indirect blood pressure determination by Doppler technique in renal hypertensive cats. Am J Physiol 237, H720–H723.

Mark JB, Slaughter TF (2005) Cardiovascular monitoring. In: Miller’s Anesthesia (6th edn). Miller RD (ed.), Elsevier, Churchill Livingstone, pp. 1268–1286. Martinez EA, Wagner AE, Driessen B et al. (2005) Guidelines for Anesthesia in Horses. http://www.acva.org/ diponly/action/Guidelines_Anesthesia_Horses_041227. htm. Meurs KM, Miller MW, Slater MR (1996) Comparison of the indirect oscillometric and direct arterial methods for blood pressure measurements in anesthetized dogs. J Am Anim Hosp Assoc 32, 471–475. Navqvi NH (1998) Who was the first to monitor blood pressure during anaesthesia? Eur J Anaesthesiol 15, 255–259. Nicholson A, Watson A (2001) Survey on small animal anaesthesia. Aust Vet J 79, 613–619. Paddleford RR (1992) The case for routine intraoperative blood pressure monitoring. Vet Clin North Am Small Anim Pract 22, 444–445. 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. Sawyer DC, Guikema AH, Siegel EM (2004) Evaluation of a new oscillometric blood pressure monitor in isofluraneanesthetized dogs. Vet Anaesth Analg 31, 27–39. Stepien RL, Rapoport GS (1999) Clinical comparison of three methods to measure blood pressure in nonsedated dogs. J Am Vet Med Assoc 215, 1623–1628. Tibby SM (2002) Measurement of cardiac output and tissue perfusion. Curr Opin Pediatr 14, 303–309. Waelchli-Suter CM, McDonell WN, Pascoe PJ (1986) Evaluation of an Esophageal Stethoscope as an early predictor of serious cardiovascular insufficiency in the dog. Vet Surg 15, 453–457. Wagner AE, Brodbelt DC (1997) Arterial blood pressure monitoring in anesthetized animals. J Am Vet Med Assoc 210, 1279–1285. Wagner AE, Hellyer PW (2002) Observations of private veterinary practices in Colorado, with an emphasis on anesthesia. J Vet Med Educ 29, 176–182. Wagner AE, Wright BD, Hellyer PW (2003) Myths and misconceptions in small animal anesthesia. J Am Vet Med Assoc 223, 1426–1432. Webb RK, Van Der Walt JH, Runciman WB et al. (1993) Which monitor?An analysis of 2,000 incident reports. Anaesth. Intensive Care 21, 529–542. Weiser MG, Spangler WL, Gribble DH (1977) Blood pressure measurement in the dog. J Am Vet Med Assoc 171, 364–368. Young SS (1999) Anesthesia monitoring systems. http:// info.med.yale.edu/yarcs/vcs/monitori.htm. Received 7 March 2006; accepted 28 December 2006.

 2007 The Authors. Journal compilation  2007 Association of Veterinary Anaesthetists, 35, 13–21

21