Materials and Methods

The procedures performed in this study conformed to the Guiding principles in the care and use of animals approved by the American Physiological Society.

All client-owned dogs and cats admitted to the University of California Veterinary Medical Teaching Hospital small animal ICU from October 2004 to June 2005 were considered for the study. Dogs and cats with a functional IV jugular catheter and 1 or more peripheral IV catheters were included in the study. A central catheter was considered to be functional if blood could be flushed and aspirated easily, and a peripheral catheter was considered to be functional if the catheter could be flushed easily. Animals were excluded if 1 or both catheter sites developed signs of inflammation or infection, if the typical sinusoidal CVP waveform was not consistently present when the catheter was connected to the pressure monitor, or if there was radiographic evidence of space-occupying intrathoracic disease such as pleural effusion or an intrathoracic mass. All catheters were placed by the primary clinician for the purpose of patient care.

Peripheral venous catheters were placed in the cephalic, lateral saphenous, and medial saphenous veins. Peripheral IV catheters^a ranged in size from 18 to 22 gauge in dogs and 20 to 22 gauge in cats. The 22-gauge catheters were available in lengths of 3, 4, and 20 cm. The 3- and 4-cm catheters were combined into 1 group for analysis. In patients with > 1 peripheral catheter, each catheter was evaluated. Triple-lumen^b or quadruple-lumen^c 5- to 8.5-F central IV catheters were introduced through the left or right jugular vein and extended into the cranial vena cava. The distance from the catheter insertion site to the caudal aspect of the shoulder was measured before each catheter was placed to ensure appropriate placement of the tip of the central venous catheter. Measurements from multilumen catheters were made from the distal port.

In each dog or cat, central and peripheral venous catheters were connected via a 3-way stopcock, high-pressure tubing, and pressure transducer^d to a single blood pressure monitoring system.^e The monitor displayed both pressure tracings simultaneously and in real time and calculated mean pressure to the nearest 1 mm Hg by use of a built-in algorithm. The transducer, tubing, and catheter systems were flushed continuously with saline (0.9% NaCl) solution and heparin solution (4 U of heparin/mL) at a rate of 3 mL/h. Pressure transducers were taped side by side to a stable support, placed approximately at the level of the right atrium, zero-calibrated to atmospheric pressure, and flushed. Positioning of the transducers was assessed visually. Absence of catheter obstruction was determined by monitoring for an appropriate pressure waveform. A square wave test was performed to verify appropriate system damping.¹² Peripheral catheter measurements were performed after equilibration of the pressure reading, which was defined as absence of fluctuation in PVP > 3 mm Hg during a 5-minute period. No measurements were taken during rapid administration of IV fluid or blood products. Five simultaneous CVP and PVP measurements were obtained 1 to 2 minutes apart during a 5- to 10-minute period, and mean values were calculated.

All animals were assessed while in a resting position. Positions were recorded as left or right lateral recumbency, sternal recumbency with weight distributed evenly on all limbs (full sternal), sternal recumbency with left hip down and pelvic limbs extended to the right (left sternal), or sternal recumbency with right hip down and pelvic limbs extended to the left (right sternal). Additional information collected included species, sex, body weight, and whether the animal was being mechanically ventilated.

Statistical analysis—Analyses were performed with commercial software.^f,g Descriptive data were expressed as mean ± SD. Analyses were performed separately for dogs and cats. The mean difference between PVP and CVP was calculated. Least-squares multiple linear regression was used to model the relationship between CVP and the following explanatory variables: PVP, body position, catheter size, and peripheral catheterization site. Interactions between PVP and other explanatory variables were screened by use of t tests. Values of P ≤ 0.05 were considered significant.

Results

Thirty-six dogs representing 23 breeds were enrolled in the study, ranging in body weight from 1.4 to 63.6 kg. One dog had 2 peripheral catheters; therefore, 37 pairs of catheters were analyzed. For the peripheral catheters, 32 were positioned in the cephalic vein and 5 were positioned in the lateral saphenous vein. Four dogs were anesthetized and mechanically ventilated, and the remaining dogs were awake or recovering from anesthesia.

Ten cats representing 3 breeds were also enrolled, ranging in body weight from 2.5 to 9 kg. Two cats had 2 peripheral catheters each, and data from each catheter pair were included in analyses. None of the cats were mechanically ventilated. Because of the small number of cats enrolled, data for left and right lateral recumbency positions were combined into 1 group, as were data for left and right sternal positions. Demographic data for both species were summarized (Table 1). The distribution of peripheral catheter positions, peripheral catheter sizes, and body positions was also summarized (Table 2).

In dogs, mean PVP was 8.7 ± 5.8 mm Hg (range, 0 to 23.8 mm Hg) and mean CVP was 3.1 ± 2.5 mm Hg (range, −0.4 to 8.8 mm Hg). Peripheral venous pressure ranged from −2.2 to 22.6 mm Hg greater than CVP, with a mean PVP-CVP difference of 5.7 ± 5.8 mm Hg. In cats, mean PVP was 9.8 ± 7.2 mm Hg (range, 4.0 to 26.0 mm Hg) and mean CVP was 3.8 ± 2.8 mm Hg (range, −1.0 to 6.4 mm Hg). Peripheral venous pressure ranged from −0.4 to 20.0 mm Hg higher than CVP, with a mean PVP-CVP difference of 6.0 ± 6.9 mm Hg. Results were summarized (Figures 1 and 2).

View Full Size

Association between PVP and CVP in 36 awake dogs in a study in which those pressures were compared among various catheter sizes, catherization sites, and body positions. The straight line is the line of equality and represents the line upon which data points with a perfect positive fit (R= +1) would fall. The equation of that line is y = x.

Citation: American Journal of Veterinary Research 67, 12; 10.2460/ajvr.67.12.1987

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Association between PVP and CVP in 10 awake cats in which those pressures were compared among various catheter sizes, catheterization sites, and body positions. See Figure 1 for key.

Citation: American Journal of Veterinary Research 67, 12; 10.2460/ajvr.67.12.1987

• Download Figure
• Download figure as PowerPoint slide

The overall correlation between PVP and CVP was weak in dogs (R = 0.24) and cats (R = 0.31), and multiple regression analysis revealed no significant relationship between PVP and CVP regardless of catheter size, catheter position, or patient position, in either species (Tables 3 and 4). Further analysis of the correlation between body weight and the PVP-CVP difference also failed to reveal a significant association.

All CVP tracings revealed a sinusoidal pressure waveform, whereas all of the PVP tracings had a flat pressure waveform. In most instances, once pressure had been equilibrated initially, central and peripheral pressure tracings did not vary more than 1 to 2 mm Hg over the course of data collection or during connection and disconnection of the equipment. However, there were a few exceptions. Peripheral venous pressure, as measured through a cephalic catheter, decreased suddenly from 43.6 to 4.4 mm Hg in 1 dog resting in sternal recumbency after the dog voluntarily extended the elbow joints cranially and reduced the degree of elbow flexion. In a second dog, PVP decreased abruptly from 66 to 6 mm Hg when the dog lowered its head to the floor. In 2 other dogs, PVP decreased from 17.6 to 10.4 mm Hg in 1 dog and rose from 3 to 13 mm Hg in the other, with no apparent change in head or body position and no fluctuation in CVP > 1 to 2 mm Hg.

Discussion

Results indicated that there was poor correlation between PVP and CVP measurements in dogs and cats. Stratifying the data by peripheral catheter size, peripheral catheter position, or patient position did not strengthen the correlation.

In contrast to our results, other investigators have reported the usefulness of PVP measurements in human pediatric^3,5 and adult^4,6–8 surgical patients. Investigators have consistently found good agreement across multiple patient positions, peripheral catheter sizes, and peripheral catheter locations in both the upper and lower extremities. In a prospective study³ of the relationship between PVP and CVP in 50 pediatric elective surgical patients, correlation coefficients were 0.93 during general anesthesia and 0.88 during spontaneous ventilation after surgery. Peripheral venous pressure and CVP were compared for multiple peripheral catheter sizes (18 to 22 g) in different sites (hand, forearm, and antecubital area) and had close agreement, and PVP measurements had a mean bias of 1.92 ± 0.47 mm Hg over CVP measurements. In another study,⁶ the relationship between PVP and CVP was investigated in 500 surgical patients in 5 body positions (eg, supine, lithotomy, Trendelenburg, Fowler, and prone positions) and with 2 peripheral venous catheter locations (forearm and dorsum of the hand) and 2 catheter sizes (18 and 20 gauge); in that study, the overall correlation coefficient was 0.89. Peripheral venous pressure was consistently 1.8 to 2.3 mm Hg higher than CVP, with a mean difference of 2 ± 1.8 mm Hg.

There are several possible explanations why PVP was only weakly correlated with CVP in dogs and cats in the present study. First, venous pressure is determined by the interplay of vascular resistance, venous blood volume, and cardiac output.⁹ Vascular resistance is in turn determined by vascular tone, vessel compliance, and the number of venous valves between the central and peripheral measurement sites.⁶ Compared with humans, species differences in these cardiovascular variables may have contributed to the larger PVPCVP differences observed in many of the dogs and cats in this study.

Second, although results of multiple regression analysis failed to detect a dependency of PVP-CVP difference on body position in these dogs and cats, the possibility of venous occlusion secondary to small changes in the degree of limb flexion or extension, even in animals in the same state of recumbency, cannot be ruled out. This was dramatically demonstrated in 1 dog resting in sternal recumbency, in which PVP decreased from 43.6 to 4.4 mm Hg after slight extension of the elbow joint. However, for most animals in which there were similarly small changes in limb position, no effect on PVP was seen.

Third, in animals in which the catheter diameter approached the venous diameter, the peripheral catheters themselves may have caused partial venous occlusion.¹² Vascular obstruction by the catheter could theoretically cause pressure to increase upstream from the catheter, decreasing PVP as measured at the tip of the catheter. Although the peripheral vein diameters in individual animals were unknown, this effect would be expected to be most evident in the smallest patients, in which vein diameters were commonly only slightly greater than catheter diameters, and to be less important in larger animals, where the opposite was true. However, evidence suggests that this is an unlikely explanation for the weak correlation between PVP and CVP in our dogs and cats. Body weight does not appear to influence the relationship between PVP and CVP in humans, even in children weighing as little as 8 kg.⁵ Similarly, no relationship between body weight and PVP-CVP difference was observed in the present study.

Differences in study design may be a final reason why PVP was only weakly correlated with CVP in the present study. Studies of venous pressure in humans have been performed almost exclusively in anesthetized, mechanically ventilated patients, and anesthesia is known to modify cardiac output and vasomotor tone.¹³ In contrast, none of the cats and only 4 of 36 (11%) dogs in this report were anesthetized and mechanically ventilated. Reassessment of the relationship between PVP and CVP in a larger number of anesthetized animals may be worthwhile.

Significant moment-to-moment variability in PVP, but not CVP, was evident in 4 of 36 dogs in this study. This type of variability was not observed in cats. In 2 dogs, the cause of the sudden decrease in PVP was likely attributable to relief of venous occlusion after a slight shift in limb position. However, in 3 other dogs, the sudden increase or decrease in PVP observed was not preceded by an apparent change in limb or body position or in CVP. It is probable that 1 or more of the factors mentioned previously, such as temporary fluctuations in vascular tone, may be responsible for these acute changes in venous pressure, but the exact cause remains unknown.

Several measures were taken to improve the reliability and accuracy of the CVP and PVP measurements, including ensuring catheter patency prior to obtaining pressure readings, use of identical instrumentation, and simultaneous recording of CVP and PVP pressure readings to eliminate the influence of moment-to-moment pressure variability. The potential for confounding effects was likely minimized by these precautions. In addition, attention was paid to ensure proper placement of the central venous catheter. Our standard hospital protocol is to premeasure the distance between the cervical insertion site and the caudal aspect of the scapula, with that distance indicating the depth to which the catheter should be inserted. Prior radiographic experience has confirmed that this protocol consistently results in placement of the tip of the central venous catheter in the cranial vena cava, and radiography to confirm catheter position was not deemed necessary or performed in the present study.

One weakness of the present study was that the pressure transducers were not calibrated prior to obtaining pressure measurements. It is possible that individual transducers may have malfunctioned and yielded erroneous results. However, this is unlikely to have occurred with enough frequency to explain the large PVP-CVP differences observed. Moreover, in our hospital, pressure transducers used by the anesthesia service are commonly exchanged for new transducers when patients are moved into the ICU for postoperative recovery, and clinically important differences in pressure measurements are rarely observed.

Flat peripheral pressure waveforms were observed in all dogs and cats, consistent with the presence of increased damping of the CVP waveform when measured peripherally. The flat waveform prevented verification of the presence of a noninterrupted column of blood between the peripheral and central venous measurement sites. Both sinusoidal and flat PVP waveforms have been observed in humans,^3,4 and the presence of a flat PVP waveform does not reduce the strength of correlation with CVP. Whether this is true in animals is unknown.

Measurements of PVP in dogs and cats were not well correlated with measurements of CVP under stable hemodynamic conditions across all patient positions, catheter sizes, catheterization sites, and body weights. Measurements of PVP should not be used to approximate CVP in awake dogs and cats. The possibility that it may be a useful technique in anesthetized animals warrants further investigation.