• Open Access

Repeatability, Reproducibility, and Effect of Head Position on Central Venous Pressure Measurement in Standing Adult Horses


  • This work was performed at New Bolton Center, University of Pennsylvania School of Veterinary Medicine. Portions of these data were presented previously at the ACVIM Forum, 2009.

Corresponding author: Rose Nolen-Walston, Department of Clinical Studies, New Bolton Center, University of Pennsylvania, 382 West Street Road, Kennett Square, PA 19348; e-mail: rnolenw@vet.upenn.edu.


Background: Central venous pressure (CVP) is a used as an estimation of intravascular volume status in various species. Techniques for measuring CVP in horses have been described, but the repeatability of these readings at a single time point or over time has not been established.

Hypothesis: That CVP measurements in healthy adult horses would be repeatable at each time point, that these readings would be reproducible over time, and that alteration in head position relative to the heart would alter CVP.

Animals: Ten healthy adult research horses.

Methods: In an experimental study, horses were instrumented with a central venous catheter. Readings were taken in triplicate q6h for 2 days by water manometry, and twice daily with the head in neutral, elevated, and lowered positions by electronic manometry.

Results: Variation in the “neutral” measurements obtained at each time point was <0.1 ± 1.0 cmH2O (P= .718). There was a significant decrease in CVP over time (P= .015), which was eliminated when results were controlled for acute decrease in body weight of −1.35% (presumed hypohydration because of lack of acclimatization and decreased water intake). Head height had a significant and directional effect on CVP in that the elevated head position decreased CVP −2.0 ± 6.5 cmH2O (P < .001) while the lowered head position increased CVP by 3.7 ± 5.5 cmH2O (P < .001).

Conclusions and Clinical Importance: CVP values obtained by water manometry were repeatable in adult horses, but were reproducible only when controlled for changes in hydration. Care should be taken to maintain consistency in head position to prevent erroneous readings.


central venous pressure


early goal-directed therapy

Central venous pressure (CVP), a measurement of right atrial or cranial vena cava pressure, is an important estimation of preload in intensive care medicine. CVP can be affected by changes in intrathoracic pressure, cardiac failure, and most commonly by changes in systemic volume status.1 CVP, along with mean arterial pressure and central venous oxygen saturation, are used in early goal-directed therapy (EGDT), which has been shown to significantly decrease in-hospital mortality rates when applied to human emergency room admission with septic shock.2 Obtaining an adequate CVP is the first step in cardiovascular stabilization in EGDT, and aids in directing fluid therapy and monitoring patients with ongoing losses. The measurement of CVP has been used in studies exploring effective resuscitation fluids in dogs with gastric dilatation volvulus,3 and decreases in horses with acute hemorrhage4 and chronic dehydration.a

In order to use this technique as part of an equine- specific EGDT protocol, accurate measurements are essential. CVP measurements are typically obtained from horses by water manometry.4–6 While the validation of the placement of central venous catheters has been established,7 the temporal stability of the measurements and the effects of changes in head elevation have not been validated. Because of their relatively long necks and variety of normal head positions while standing, CVP values in standing horses might show significant inconsistencies, especially when compared with humans and small animal species, where CVP is typically measured in recumbency. In humans, variation in CVP is seen when the patient moves from supine to standing position8 and there are significant changes when anesthetized horses are moved from lateral to dorsal decubitus.6

The goal of this study was to describe the repeatability (agreement of measurements taken under identical conditions, where any variation can be assumed to be because of errors in the measurement process) and reproducibility (the variation in measurements made on a subject under changing conditions, such as time)9 of CVP measurements in standing adult horses. Additionally, we sought to identify the effects of head position (high, low, and neutral) on CVP. We hypothesized that CVP measurements obtained via water manometry in adult horses would provide repeatable results at each time point, and that these readings would be reproducible over time (days). Furthermore, we hypothesized that alterations in head position relative to the heart would significantly alter CVP.

Methods and Materials

Ten healthy adult horses from the university teaching heard were used in this nonterminal study. Horses were between the age of 7 and 14 years, weighing between 450 and 700 kg, and were Thoroughbreds, Standardbreds, and Warmbloods. During these experiments the horses were allowed access to ad lib hay and water and were fed grain twice daily. The amount of water consumed was recorded every 6 hours. Physical parameters (heart rate, respiratory rate, temperature, borborygmi, and manure production), PCV, total protein, and weight were recorded every 12 hours. All procedures were approved by the Institutional Animal Care and Use Committee at the University of Pennsylvania.

A 70-cm 16 g single lumen central venous catheterb was placed in the right jugular vein at the junction of the proximal third and middle third of the neck. The catheter was advanced ∼60 cm. Appropriate placement was confirmed by small oscillatory movements of the indicator ball in a water manometerc connected to the system that corresponded to changes in intrathoracic pressure during the respiratory cycle.4,6 An indicator line was clipped over the point of the right shoulder as a reference point for the height of the right atrium6 as a zero point for CVP readings.

Experiment 1

With the horse standing lightly restrained with head in a neutral position, the water manometer was directly connected to the central venous catheter by removing the injection cap. A 60-cm3 syringe filled with heparinized saline was connected to the 3-way stopcock and the water manometer filled to 25 cm. With the “zero” of the water manometer held at the point of the shoulder the manometer was opened to the venous circulation. The reading was taken when the indicator ball was no longer dropping but moving synchronously with the animal's respiration. This process was repeated in immediate succession 3 times to assess repeatability, and the means of these triplicate readings obtained every 6 hours were used to assess reproducibility over 48 hours.

Experiment 2

To compare the readings at various head positions, measurements were obtained using 2 electronic monitorsd,e connected to the central venous catheter using an electronic transducerf to facilitate consistency of measurements at each head position. With the 3-way stopcock closed to the circulation, the monitor was “zeroed” as per the manufacturer's instructions. This monitoring system provides a pressure tracing as well as a numeric value. The electronically averaged measurements generated by the monitor were recorded when the pressure tracing displayed reflected the normal respiratory oscillations; no effort was made to time the readings with the respiratory cycle. Three measurements were recorded in rapid succession at each position, and a 30-second delay was allowed for pressure stabilization between each change in position. Readings were first obtained with the horse's head in neutral position (muzzle level with the point of the shoulder). The head was then manually elevated so that the most rostral aspect of the head was level with the withers, then the horse's head was lowered by allowing them to eat grain from the ground. All measurements were taken with a single device, and then the procedure repeated with the 2nd device. Readings were taken twice a day for 48 hours.

Statistical Analysis

The repeatability, reproducibility, and head positioning data were analyzed by mixed effects modeling.g The normality of the data were verified and statistical significance was set at P < .05. Data are reported as ± 1 standard deviation.


Experiment 1

Mean CVP in the neutral head position was 9.4 ± 3.6 cmH2O. In regards to repeatability, a mean of each set of 3 repetitions was obtained, and each value's difference from the mean calculated. The average difference from the mean for the 1st, 2nd, and 3rd readings were 0.07 ± 0.89, −0.01 ± 0.68, and −0.06 ± 0.8 cmH2O, respectively (Fig 1). No significant difference (P= .718) was demonstrated between repetitions at any time point. There was a significant difference between CVP measurements obtained on day 1 versus days 2 and 3 (P= .002 and .015). A decline in body weight of −1.35% (mean, range: −0.35 to −2.39%) occurred in all animals over the 2-day period of the study (Figure 2). Over the first 48 hours, water consumed by drinking was a mean of 41.2 mL/kg/d, which is considerably less than published normal values (64.4 mL/kg/d total water intake,10 >90% of which is obtained via drinking). The model was thus expanded to control for the confounding influence of change in weight, whereupon no significant difference was found between CVP values at baseline and subsequent days (P= .359 and .237).

Figure 1.

 Repeated measurement at a single time point. The y-axis shows the difference of the value of each individual repetition (1, 2, 3 on x-axis) from the mean of the 3 central venous pressure (CVP) measurements taken in immediate succession (n = 210 total CVP measurements).

Figure 2.

 Mean body weight (BW) and central venous pressure (CVP) on day 1 versus day 3. This graph demonstrates a reduction of body weight (± SE) over the 48-hour period of the study, which corresponds to a decrease in CVP over time. Upper and lower borders of the boxes represent the 25 and 75% quartiles. The ends of the whiskers show the minimum and maximum values and the median value is described by the central band.

Experiment 2

Head height had a significant and directional effect on CVP. Using an electronic transducer, the mean CVP was 9.56 ± 4.2 cmH2O. The elevated head position decreased CVP by a mean of −2.0 ± 6.5 cmH2O (P < .001) while the lowered head position increased CVP by 3.7 ± 5.5 cmH2O (P < .001) (Fig 3).

Figure 3.

 The effect of head position on central venous pressure. The plots represent the pressure measurements with the head in lowered (down), neutral (neutral), and elevated (up) head position. Measurements were obtained using two electronic pressure wave monitors, Datascope Passport LTd (p) and Medtronic Lifepak 12e (m). Lowering and elevating the head had a significant effect on central venous pressure as compared with neutral position.


This study demonstrates excellent repeatability (agreement of measurements taken under identical conditions) of CVP measurements obtained by water manometry in healthy adult horses. Although raw measurements over time did not show adequate reproducibility (the variation in measurements made on a subject under changing conditions, such as time), this finding is probably explained by changes in hydration status during this period. Variations in head position altered the readings and led to erroneous measurements of CVP.

Clinically, at the authors' institution, readings are routinely taken in triplicate to avoid a single reading giving erroneous results. However, this study shows that there is <2% variability among repeated measures, indicating that in euvolemic horses, a single measurement may be adequate. Not only were the results repeatable, but the measurements obtained were consistent with previously reported reference range of 7.5 ± 0.9–12 ± 6 cmH2O.4,6,7 The proven repeatability of measurements should encourage confidence in the use of water manometry as a consistent measure of monitoring volume status in adult equids. In addition to providing consistent results, the central venous catheters were easy to place under minimal restraint. All horses tolerated the catheter placement and repeated readings without incident. The catheters remained indwelling for a total of 4 days throughout the course of this study and no evidence of superficial infection or thrombophlebitis was noted.

In humans, one of the main sources of error is variability in the level chosen for the “zero” mark of the manometer.11 This problem was alleviated in this study by the use of a small clipped line to ensure consistent height of the manometer “zero,” and should be recommended for clinical cases. Obtaining readings during end expiration is considered ideal in obtaining accurate values in humans12; a paradoxic elevation of CVP during inspiration is referred to as Kussmaul's sign, and is classically associated with constrictive pericarditis.13,14 In this study, although respiratory oscillations were noted during measurement by movement of the indicator ball in the water manometer, they were rarely more than approximately 0.5 cmH2O and were not considered clinically significant. With the electronic monitors used there was mild oscillation of the pressure wave associated with respiration and the electronically generated average displayed occasionally changed with respiration, but again there was rarely a change >0.5 cmH20. In cases of increased respiratory effort or mechanical ventilation, it would be easier to identify the exact point of end expiration, but in both cases CVP values may differ from horses spontaneously breathing normal tidal volumes.

Initial analysis suggested that CVP in healthy adult horses had poor reproducibility as a function of time. It is unlikely that horses lost body mass because of reduction in muscle or fat tissue over such a short period while eating an adequate diet, although changes associated with defecation and urination might also cause fluctuations in body weight. However, it was noted that while the animals all appeared to remain clinically well hydrated, there was a consistent drop in body weight over the course of the study. This was most likely because of a lack of acclimatization before the start of and poor water intake during the study. The horses, normally housed on lush grass, were stabled and fed a diet with lower water content, and had water intake that was below maintenance requirements. Previous studies have shown that for every percentage decrease in body weight because of dehydration, there is a 2.2 cmH2O decrease in CVP.a Supporting this hypothesis, analysis of the effect of the interaction of weight by day removed the deviance in the CVP measurements; specifically the mean decrease of 1.35% body weight could account for a decrease of 2.97 cmH2O, which approximates the change seen in the horses (Fig 3). Further investigation regarding the reproducibility of CVP measurements in horses is warranted to validate this conclusion.

Head position proved to be extremely important in obtaining consistent readings. This portion of the study was performed using 2 electronic pressure monitors, which have been validated for use in horses.h Readings obtained by this method in the neutral position also fall within previously published reference ranges.6 Prehension, mastication, and swallowing of grain did not appear grossly to affect the pressure tracing or of the readings obtained with the head lowered. However, significant deviation from baseline value occurred in both the head elevated and the head lowered position. These findings are consistent with several studies published in both the human and veterinary medical literature regarding positional and orthostatic effects on CVP. Although the effect of change in body position is fairly small in healthy humans,13 the effect can be profound in patients with heart failure15 or volume depletion, where there are decreases of up to 10 cmH2O when position is changed from supine to semiupright.16 This effect is used clinically in patients undergoing liver resections, who are placed in a reverse-Trendelenberg (head elevated) position to intentionally lower CVP and decrease bleeding during surgery.17 While the changes in head position in the horse may alter the CVP in the same direction as in man, it is likely that the physiologic cause of the change is different. In man, when the legs are elevated there is greater venous return from the caudal half of the body leading to an increase in pressure in the vena cava,18 and no change is seen in compliance of the vessel after even after 90 minute head-down tilt.19 In the horse, when the head and neck is lowered, it can be theorized that there is causing a pooling of blood in the external jugular and cranial vena cava leading to an increase in measured CVP.

In conclusion, this study shows that readings obtained by this method of water manometry are repeatable, may be reproducible if controlled for hydration (as determined by change in body weight), and are consistent with previously published values. We also demonstrate that it is important to perform all readings with the head in a standardized position to prevent erroneous measurements, as raising and lowering the head causes a respective decrease and increase in CVP.


a Nolen-Walston RD, Norton JL, Boston R, et al. The effects of dehydration on central venous pressure in adult horses. J Vet Intern Med 2009;23:777 (abstract)

b PICC, Peripherally Inserted Central Catheter Set, Arrow International, Reading, PA

c Central Venous Pressure Manometer, Smiths Medical ASD Inc, Dublin, OH

d Datascope Passport EL, MAQUET Cardiovascular, Wayne, NJ

e Medtronic Lifepak 12, Physio-Control Inc, Redmond, WA

f Transpac IV, Hospira, Lake Forest, IL

g Stata 10.0, StataCorp, College Station, TX

h Norton JL, Boston R, Underwood C, et al. Comparison of water manometry to two commercial electronic pressure monitors for central venous pressure measurement in the horses. J Vet Intern Med 2009;23:777 (abstract)


Support: Funding for this project was provided by the Raymond Firestone Trust and the Arabian Horse Foundation.

Conflict of interest: None.