Comparison of Thames Medical CAT+ Doppler and SunTech Vet 20 oscillometric devices for non-invasive blood pressure measurement in conscious cats

Abstract

Objectives

A comparative assessment of systolic blood pressure (BP) measurement agreement and precision in two commonly used non-invasive BP devices was carried out in conscious cats.

Methods

Systolic BP measurements were obtained from 50 conscious cats as part of their clinical investigations. All measurements were taken by the same operator and were performed according to the American College of Veterinary Internal Medicine (ACVIM) consensus guidelines. The same cuff location and cuff size were used for paired measurements. The order of device use was randomised, and an arousal score was assigned during each procedure. Precision was assessed using standard deviation and coefficient of variance comparisons. Agreement was assessed using graphical and statistical comparisons of derived ‘delta mean’ and ‘mean of means’ data.

Results

A total of 50 cats aged between 3 months and 15 years were enrolled (29 domestic shorthair, seven domestic longhair, three British Shorthair, two Ragdoll, two Sphynx, two Persian, one Siamese, one Burmese, one Russian Blue, one Maine Coon and one Oriental) with a body weight in the range of 1.2–6.2 kg. BP measurements were in the range of 95–179 mmHg. Oscillometric measurements were associated with lower arousal scores. The Doppler device demonstrated superior repeatability precision. Body weight, sex, cuff size, cuff location, arousal score and the order in which the devices were used had no significant effect on precision or agreement. Correlation between the two devices was 0.0837 and agreement was considered clinically acceptable (<10 mmHg) in 32/50 (64%) cats. Of the 18 cats with suboptimal agreement, the oscillometric mean BP was higher in 14 (78%) cats.

Conclusions and relevance

Based on this study sample, suboptimal agreement between paired CAT+ Doppler and SunTech Vet20 oscillometric BP measurements in 36% of conscious cats suggests that these devices should not be used interchangeably. Compared with Doppler, oscillometric measurements were consistently higher, particularly at higher blood pressures. Further studies are required to assess which device has superior accuracy in conscious cats.

Introduction

Hypertension is defined as a persistent elevation in systemic arterial blood pressure (BP). In cats, the measurement of systolic pressure is considered to be more reliable than diastolic and mean pressure, and it is systolic pressure that is predominantly used to guide treatment recommendations.^1,2 According to the International Renal Interest Society (IRIS) and the American College of Veterinary Internal Medicine (ACVIM) guidelines, cats with a systolic BP >160 mmHg are considered hypertensive, and those falling into the 140–159 mmHg, 160–179 mmHg and >180 mmHg BP categories are considered to have a low, moderate and severe risk of target organ damage respectively. ³ The reported incidence of hypertension is approximately 13% in cats aged >9 years, increasing to <87% in cats with concurrent disease^1,4,5 and, since this is associated with significant morbidity, early detection and intervention is a priority, particularly in at-risk populations.^1,6,7 Standardised guidelines for the evaluation, monitoring and treatment of hypertension are available; however, the success of these protocols is reliant upon the use of reliable BP measurement techniques.^3,8

Direct BP measurement via arterial catheterisation is considered the gold standard but requires experience and specialist equipment, making it impractical for use in most clinical situations.^3,9,10 Non-invasive indirect methods using Doppler or oscillometric devices are predominantly used in clinical practice.^10,11 While Doppler methodology is still widely used in cats, oscillometric methods have been gaining favour and, most recently, high-definition oscillometry (HDO), which has the advantage of providing rapid and sensitive measurements alongside visual control and pulse wave analysis, has been promoted.^2,12 Nevertheless, previous assessments regarding the utility of the available devices have been discordant.^2,13–18 Indeed, since precision and accuracy are likely to differ depending on patient factors such as species, posture, consciousness, cuff size and choice of artery,^{19 –21} discordant results arising from differing study design are not unexpected.

While Doppler devices all use the same methodology, oscillometric devices use unique manufacturer-designed algorithms to calculate automated systolic, mean and diastolic arterial pressure values,^15,22 thus each oscillometric device requires individual validation. The SunTech Vet20 is a commonly used traditional (non-high-definition) oscillometric device with an animal-specific algorithm. A previous study has investigated its use in anaesthetised cats ²³ but, with the exception of a short white paper published by the manufacturer, ²⁴ there is no information available regarding its use in conscious cats. Since it has been developed specifically for the veterinary market and is promoted as being ideal for conscious patients (quick to use, quiet and motion tolerant), the aim of this study was to evaluate this particular oscillometric device in conscious cats via comparison with the widely used CAT+ Doppler device.

Material and methods

Ethical approval was granted by the Royal College of Veterinary surgeons (RCVS) ethical approval committee. A total of 50 client-owned cats were prospectively enrolled between April 2021 and November 2021. Inclusion required the need for BP measurement as part of a patient’s diagnostic workup. Cats not requiring BP measurement and cats under sedation or general anaesthesia were excluded. Stress was minimised by acclimatising the patient to a quiet environment and using only gentle restraint. The procedure was abandoned and the cat eliminated from the study, or reassessed at a later date, in the event of excessive movement or stress-related behaviour being displayed during the use of either device. Body weight, age, sex, breed, level of arousal, cuff size, cuff location and order of device use were recorded for each patient.

BP measurements were performed according to the ACVIM consensus guidelines using a cuff size of 30–40% of the limb circumference. ³ Both devices were operated according to the manufacturer’s guidelines. All measurements were obtained by a single operator, and in individual patients, the same cuff size and location were maintained for both devices. Mean systolic BP values were calculated using 5–10 measurements after discarding the initial outlier results as recommended by the ACVIM consensus guidelines. All measurements were taken sequentially with the order of use of devices being assigned randomly. The level of patient arousal during each procedure was recorded using a scoring system similar to that previously suggested by Scansen et al ²⁵ and adapted as follows:

• Level 1 = measurements can be performed with ease by one investigator alone and the patient remains still;

• Level 2 = measurements can be performed by one investigator alone, but the cat shows movements, such as an occasional limb movement or shifting of body position;

• Level 3 = measurements can be performed by one investigator alone, but have to be repeated due to movements of the cat;

• Level 4 = a second person has to restrain the cat in order to perform the procedure, but measurements are then taken with ease;

• Level 5 = a second person is required to restrain the cat in order to perform the procedure and measurements have to be repeated due to movements of the cat.

Statistical analysis

A sample size calculation, assuming a systematic error of <5 mmHg, and an uncertainty of σ/√ N revealed that the enrolment of 50 cats would facilitate the detection of a 5 mmHg difference with 5σ confidence. Following the assessment of Q-Q plots and a Kolmogorov–Smirnov Goodness of Fit Test, body weight, mean BP and delta mean BP (the latter defined as Doppler mean BP minus oscillometric mean BP) were examined as parametric data. The results were presented as mean ± SD, and P <0.05 was considered statistically significant. Age, absolute delta mean (the non-directional equivalent of delta mean), SD and coefficient of variance (CV) were assessed as non-parametric continuous variables and were presented as median and interquartile range (IQR). Mean BP, delta mean BP and age were also examined as categorical variables, as were sex, arousal score, cuff size and cuff location. In order to maintain roughly equal patient numbers in each group, mean BP, delta mean BP, age and bodyweight were each categorised into three groups as follows: <115 mmHg, 115–135 mmHg and >135 mmHg; <5 mmHg, 5–10 mmHg and >10 mmHg; <8 years, 8–13 years and >13 years; and <3.5 kg, 3.5–4.5 kg and >4.5 kg. Due to the small group sizes, breed category was not included in the statistical analysis.

To assess precision, the mean BP standard deviation and CV data were compared using the related-samples Wilcoxon signed-rank test. The Kruskal–Wallis test was used to assess precision between mean BP categories. The agreement between continuous Doppler and oscillometric BP data was examined using scatter plots, Bland–Altman plots and Pearson correlation. The relationship between mean and delta mean data was demonstrated using box and whisker plots and Pearson correlations. Kruskal–Wallis and one-way ANOVA tests were applied for univariate analysis of categorical variables with non-parametric and parametric data, respectively. Multivariate ANOVA was applied for multivariate analysis of mean BP data with categorical variables. All statistical analyses were performed using the SPSS software package (SPSS 10.0 for Windows; IBM).

Results

Study population

A total of 50 cats were enrolled in the study, comprising the following breeds: domestic shorthair (n = 29); domestic longhair (n = 7); Ragdoll (n = 2); Sphynx (n = 2); Persian (n = 2); Russian Blue (n = 1); Maine Coon (n = 1); Oriental (n = 1); Siamese (n = 1); and Burmese (n = 1). The study sample included 30 male neutered cats, 16 female neutered cats, two female entire cats and one female cat with unknown neuter status. The median age and weight were 9.3 years (range 3 months–15 years) and 4.32 kg (range 1.2–6.2 kg), respectively.

Blood pressure readings

A total of 10 BP readings were taken with each device in 48/50 cats. In the other two cats, seven and eight measurements were taken with the oscillometric device, respectively, and 10 with the Doppler device. BP measurements were in the range of 95–173 mmHg with the Doppler device and 96–179 mmHg with the oscillometric device. The Doppler device was used first in 19/50 cats. The cuff size was in the range of 1–3 and was applied to the tail in 20 cats, the right forelimb in 11 cats and left forelimb in 19 cats.

Arousal score

Using the Doppler device, 30 (60%) cats had an arousal score of 1, 14 (28%) had a score of 2, 1 (2%) had a score of 3 and 5 (10%) had a score of 4. No cats had a score of 5. Using the oscillometric device, 40 (80%) cats had a score of 1, 8 (16%) had a score of 2 and 2 (4%) had a score of 4. No cats had a score of 3 or 5 (Figure 1). In 12/50 cats, paired arousal scores differed between the two techniques, with the Doppler device being associated with the higher score on each occasion, regardless of the order of use (Table 1). In 6/12 (50%) cats, good mean BP agreement (<10 mmHg) was maintained between both devices. Of the 6/12 with poor BP agreement (>10 mmHg), the higher arousal score correlated with the higher mean BP in only two cats (Table 1).

Figure 1.

Comparative (a) Doppler and (b) oscillometric arousal score data [arousal score (no. of cats, % cats)]

Table 1.

Paired mean and delta mean blood pressure (BP) data for cats with disparate arousal scores

Cat ID	Arousal score (Doppler)	Arousal score (SunTech)	Mean BP (Doppler)	Mean BP (SunTech)	Delta mean (Doppler – SunTech)
2	2	1	115	112	3.3*
5	2	1	173	174	−0.8
8	3	2	153	180	−26.0 ‡
10	2	1	121	138	−17.7 ‡
19	2	1	115	112	3.6*
20	4	1	121	144	−23.4 ‡
24	2	1	119	122	−3.1*
26	4	2	125	109	−16 †
31	2	1	98	96	1.8*
34	2	1	117	120	−2.6
46	2	1	149	180	−30.3 ‡
47	4	1	129	115	14.0 †

Level of BP agreement:

^*Delta mean <10 mmHg

^†Delta mean 10–20 mmHg

^‡Delta mean >20 mmHg

Precision

Subjective scatter plot comparisons of paired mean BP with standard deviation and CV suggested superior repeatability precision of the Doppler device (Figure 2). Statistical analysis confirmed a significant difference between paired mean BP standard deviation (P = 0.025) and between paired mean BP CV (P = 0.0044) (Table 2). SD and CV comparisons between categorised mean BP groups were non-significant for both the Doppler (SD P = 0.100; CV P = 0.796) and oscillometric (SD P = 0.193; CV P = 0.416) devices (Table 2).

Figure 2.

(a,b) Doppler and (c,d) oscillometric precision illustrated by (a,c) SD and (b,d) coefficient of variance data (red lines = mean ± SD)

Table 2.

Doppler vs oscillometric and within-device BP category precision comparisons

	n	Median	IQR	95% CI	DoF	P value
Doppler vs oscillometric
Doppler SD	50	3.353	2.20	3.034–3.852	49	0.025
SunTech SD	50	4.057	4.84	3.958–6.225
Doppler CV	50	0.026	0.02	0.025–0.031	49	0.044
SunTech CV	50	0.032	0.03	0.031–0.049
Within-device BP
Doppler SD <115 mmHg	16	2.792	2.10	2.276–3.395	2	0.100
Doppler SD 115–135 mmHg	20	3.416	2.33	2.854–4.356
Doppler SD >135 mmHg	14	3.769	2.22	3.082–4.733
SunTech SD <115 mmHg	16	2.582	2.87	2.022–5.419	2	0.193
SunTech SD 115–135 mmHg	15	4.037	4.88	3.369–7.578
SunTech SD >135 mmHg	19	5.280	5.01	3.735–8.155
Doppler CV <115 mmHg	16	0.0261	0.02	0.022–0.033	2	0.796
Doppler CV 115–135 mmHg	20	0.0292	0.02	0.023–0.031
Doppler CV >135 mmHg	14	0.0251	0.01	0.021–0.031
SunTech CV <115 mmHg	16	0.0235	0.03	0.018–0.053	2	0.461
SunTech CV 115–135 mmHg	15	0.0341	0.04	0.028–0.063
SunTech CV >135 mmHg	19	0.0336	0.04	0.025–0.055

BP = blood pressure; CV = coefficient of variance; DoF = XX; IQR = interquartile range

Agreement

Visual assessment of paired mean BP data via a scatter plot suggested a negative bias (SunTech mean > Doppler mean) which, subjectively, appeared worse for higher BP values (Figure 3). As a major outlier (>3rd quartile + 1.5 × IQR), cat 21 was removed from the dataset before the statistical analysis of delta mean data (Figure 3). The correlation between the two devices was good (Pearson coefficient 0.083, 95% CI 0.72–0.90, P <0.001) but substantial data point dispersion on a Bland–Altman plot implied suboptimal agreement (Figure 4). Box and whisker plots of categorised data suggested a tendency towards poorer agreement in cats with higher mean oscillometric BP readings (Figure 5). This was confirmed by a significant difference in absolute delta mean between the oscillometric mean BP categories (P = 0.019) (Table 3) and a negative correlation between delta mean BP and oscillometric mean BP (Pearson coefficient −0.451, P = 0.001). No difference was found between the Doppler BP categories (P = 0.76) and no correlation was found between delta mean and Doppler mean BP (Pearson coefficient 0.124, P = 0.397). Based on the assumption that absolute delta mean <10 mmHg would be clinically acceptable, adequate agreement was found in 32/50 (64%) cats (Figure 6), with 21/32 (66%) of these having excellent correlation (delta mean <5 mmHg). Agreement was considered to be clinically inadequate (>10 mmHg) in 18/50 (36%) cats, with 8/18 (44%) of these having very poor correlation (delta mean >20 mmHg). In 14/18 cats with inadequate agreement, the higher mean BP was associated with the oscillometric device (Figure 7).

Figure 3.

Direct comparison of Doppler and oscillometric mean blood pressure data (black line = perfect agreement, red circle = major outlier)

Figure 4.

Bland–Altman plot illustrating level of agreement between Doppler and oscillometric devices (red line = mean delta mean, yellow lines = mean ± 10 mmHg, green lines = 95% confidence limits)

Figure 5.

Comparison of paired mean blood pressure agreement based on (a) categorised Doppler and (b) oscillometric mean datasets (blue box = median ± IQR, green line = perfect agreement, orange line = best fit line)

Table 3.

Comparison of paired mean BP agreement based on categorised Doppler and oscillometric mean BP data

Mean BP category (mmHg)	Device	Number of cats	Mean deltamean ± SD	95% CI	Median delta mean ± IQR	Min-max (range)delta mean	DoF	P value
<115	Doppler	16	−5.3 ± 8.2	−9.6 to −0.9	−1.5 (15.5)	−22.1–3.3 (25.4)	2	0.602
115–135	Doppler	20	−6.7 ± 14.6	−13.7 to −0.7	−5.0 (23.0)	−35.8–16.0 (49.8)
>135	Doppler	13	−2.8 ± 13.4	−10.0–5.4	−0.8 (14.6)	−30.3–16.6 (46.9)
<115	SunTech	16	0.1 ± 8.8	−4.6–4.8	1.4 ( 8.9)	−15.2–16.0 (29.2)	2	0.008
115–135	SunTech	14	−2.4 ± 9.5	−7.8–3.1	−2.9 (12.1)	−22.1–10.2 (32.3)
>135	SunTech	19	−11.7 ± 14.2	−18.6 to −4.9	−9.8 (23)	−35.8–16.6 (52.4)

CI = confidence interval; DoF = XX; IQR = interquartile range

Figure 6.

Level of agreement between paired Doppler and oscillometric mean blood pressure measurements

Figure 7.

Paired mean blood pressure comparisons (green line = delta mean >10 mmHg, red line = delta mean >20 mmHg

Univariate analysis identified no association between mean or delta mean BP with sex, bodyweight category, order of device use, arousal score, cuff size or cuff location. Missing data (age: n = 2, bodyweight: n = 5, cuff size: n = 1) resulted in only 41 cats being included in the multivariate analysis, for which non-significant relationships were maintained. A non-monotonic relationship precluded age correlation as a continuous variable. Univariate analysis of categorised data identified a statistically significant difference between age and mean BP for both devices (Doppler P = 0.013, SunTech P <0.001), with cats in the older age categories having higher mean BP values. This relationship was not maintained in multivariate analysis independent of bodyweight, arousal score, cuff size and cuff location.

Discussion

Comparative data for the SunTech Vet20 oscillometric and the Thames Medical CAT+ Doppler BP devices in conscious cats are reported. Although previous studies have addressed a similar subject, they have comprised smaller sample sizes, have investigated different oscillometric devices and/or have involved anaesthetised cats. As per the ACVIM guidelines, it is essential that BP devices are validated in the species of interest and under the circumstances in which the patient is being tested,^3,8 thus it is not possible to extrapolate previous results to these particular devices in conscious cats.

With the exception of age, patient-related factors (sex, bodyweight, order of device use, arousal score, cuff size and cuff location) had no association with mean BP or with device agreement in our study. Although not maintained in multivariate analysis, a positive correlation between age and BP is widely documented.^{21,26 –30} Despite a lack of statistical association between cuff location and arousal score, our experience is that use of the tail location is helpful in reducing stress in cats that are less tolerant of manipulation. Similarly, a previous study found that use of the tail location was better tolerated in conscious cats and was associated with shorter procedure times and fewer failed measurements. ³¹ The lack of documentation of procedure times and failure rates is a limitation of our study but, as documented in previous studies,^{16,32 –34} despite its ease of use, the oscillometric device was subjectively associated with longer procedure times, largely due to failed measurements.

In the event of differing paired arousal scores, use of the Doppler device was invariably associated with the higher score. This finding was not dependent on the order of device use suggesting that, despite being associated with shorter procedure times, the Doppler technique was less well tolerated; a conclusion that would seem plausible given the different methodologies required for device use. Despite differing paired arousal scores, the agreement of paired mean BP remained adequate (<10 mmHg) in many of this subset of cats (6/12). Counterintuitively, in all but two of the six remaining cats, the higher arousal score was associated with the lower mean BP, a possible explanation being overestimation of BP by the oscillometric device or underestimation by the Doppler device. Indeed, a negative bias (SunTech mean BP > Doppler mean BP) was present in 74% of cats with poor paired mean BP agreement. While a study by Cerna et al ²² also reported higher comparative BP readings using oscillometric compared with Doppler methodology in conscious cats, several studies in anaesthetised cats have found the opposite, particularly at higher BPs.^16,34 In addition, Petric et al ³³ reported higher oscillometric readings at low pressures and lower readings at high pressures compared with Doppler measurements in anaesthetised cats, while Jepson et al ³² reported no consistent bias in systolic BP when comparing an oscillometric device with Doppler methodology in conscious cats. Deterioration of device agreement with increasing BP has been reported previously^9,18,34 and, in our study, this was specifically associated with the oscillometric BP measurements. The inclusion of retinal examination results would have helped to further explore this finding by aiding in the verification of true hypertension in this subset of patients; unfortunately, our study methodology did not permit this.

Comparative repeatability precision of the SunTech Vet20 and Doppler methodology has not been previously reported; however, studies evaluating other oscillometric devices in cats have variably reported equivalence ¹⁶ or, as found in our study, superiority of Doppler methodology in cats.^14,32,33 While the significant difference found between paired mean BP standard deviations in this study suggests the possibility of superior precision using the Doppler methodology, it is important to acknowledge that shorter procedure times also favour lower BP variability. If, as suspected, the oscillometric device procedure times were longer, the potential for more BP variation during this technique may have confounded any precision comparisons. While enabling standardisation of measurements, the use of only one operator for all BP measurements precluded the assessment of reproducibility precision.

When compared with direct BP methodology, Doppler measurements have generally been found to correlate more closely with mean rather than systolic pressure.^10,14,15,35 With respect to anaesthetised cats, several studies have described poor agreement with direct systolic BP measurements for both Doppler^13,35 and traditional oscillometric^36,37 devices. While others have reported good correlation for Doppler ³⁴ and traditional oscillometric devices,^15,18,38 variable and often considerable bias with wide standard deviations calls into question any clinical utility of these devices for the measurement of systolic BP. Studies designed to simultaneously compare Doppler and oscillometric techniques with direct BP measurement variably report superior accuracy of the Doppler ¹⁶ and oscillometric methodology. ¹⁴ Nevertheless, based on the comparative data, measurements using any of these devices should be interpreted cautiously in individual patients. Only three studies have evaluated the accuracy of non-invasive BP measurement in conscious cats. The first of these concluded that the Doppler technique consistently underestimated direct BP measurements. ¹⁷ While the authors suggested that the Doppler methodology was nevertheless a ‘reliable’ surrogate for direct BP measurement based on strong correlation, any true clinical assessment was precluded due to incomplete data assessment, particularly a lack of Bland–Altman analysis. A later study comparing both Doppler and oscillometric methodology concluded that the former provided a more accurate reflection of direct BP in conscious cats; however, the reported mean BP standard deviations suggested unacceptable clinical utility for either device. ³⁴ The most recent and most comprehensive study investigated an HDO device and concluded that it provided a ‘faithful measurement’ of systolic BP in conscious cats. ² Indeed, although requiring confirmation in a larger study sample with more stringent methodology, to date, this is the only non-invasive BP monitor reported to have met the validation criteria defined by the ACVIM consensus panel.

According to the ACVIM guidelines, ⁸ BP device validation requires three conditions to be met: (1) mean delta mean <10 mmHg and standard deviation <15 mmHg; (2) correlation coefficient >0.9; and (3) 50% of measurements within 10 mmHg, and 80% within 20 mmHg of the gold standard. In the previous study comparing the SunTech Vet20 with direct BP measurement in anaesthetised cats, criteria 1 and 3 were met but not 2: ²³ mean delta mean 2.7 ± 13.0 mmHg, correlation coefficient 0.63, 59.3% measurements within 100 mmHg and 84.7% within 20 mmHg. Assuming Doppler as the reference device in our study, mean delta mean was −5.2 ± 9.7 mmHg, the correlation coefficient was 0.84, 62% of measurements were within 10 mmHg and 84% of measurements were within 20 mmHg. Thus, as found in anaesthetised cats, conditions 1 and 3 were met but not 2. Incomplete published data from the SunTech Medical pilot study of 29 conscious cats precludes a comprehensive comparison; however, a mean bias of −2.7 ± 28.1 mmHg was reported compared with Doppler methodology. Regardless, since no ‘gold standard’ was included in either the present study or that of the original pilot study, the choice of reference device is arbitrary. Indeed, in the absence of a validated comparative device, the assessment of accuracy was not within the remit of this study. Irrespective of this, the results suggest that the Doppler and SunTech Vet20 devices should not be used interchangeably.

In this study sample, the mean Doppler BP was in or above the upper half of the reference interval (RI; >135 mmHg) in only 14/50 cats, and only four of these would be considered hypertensive (>160 mmHg). When considering the oscillometric device, only 19/50 cats had a mean BP in or above the upper half of the RI, of which five were hypertensive. While studies establishing normal RIs have reported varied results,^{21,28,29,39 –43} most propose a mean BP of 120–130 mmHg, which correlates well with the results (Doppler mean of 124 ± 21 mmHg and oscillometric mean of 128 ± 22 mmHg) observed in this study. Nevertheless, a skew towards cats with low normal and subnormal BP is likely to have reduced the statistical power with regard to the assessment of precision and agreement at higher BPs.

Conclusions

Based on this study sample, suboptimal agreement between paired CAT+ Doppler and SunTech Vet20 oscillometric BP measurements in 36% of conscious cats suggests that these devices should not be used interchangeably. Compared with Doppler, oscillometric measurements were consistently higher, particularly at higher BPs. Further studies are required to assess which device has superior accuracy in conscious cats.