Factors influencing the development of jugular thrombophlebitis in cattle and comparison of 2 types of catheter

Abstract

Two studies were conducted to evaluate the factors associated with the development of thrombophlebitis in cattle following intravenous catheterization of the jugular vein. In study 1, 20 healthy animals were catheterized with 2 different types of catheter (polytetrafluoroethylene (PTE) versus polyvinylchloride (PVC)) for a period of 120 hours. In study 2, 50 dairy cows referred for treatment to the Large Animal Veterinary Teaching Hospital at the University of Montréal were catheterized with a commercial PTE catheter, using a standardized technique, for a period varying from 3 to 6 days, and inherent risk factors were evaluated. A clinical and echographical evaluation of the 2 groups in study 1 demonstrated a higher frequency of thrombosis in the PTE group than in the PVC group. In study 2, the primary factors associated with the development of thrombophlebitis in sick cows were the experience of the manipulator and the severity of the disease.

Introduction

With the increased value of animals and availability of therapeutic agents, IV administration of drugs and fluids is often a crucial factor in the treatment of different field conditions. Intravenous catheterization is indicated in order to decrease the frequency of thrombophlebitis when repetitive IV injections are required and also to facilitate treatment by the owners. However, the utilization of IV catheters has inherent risks, with the physicochemical properties of the catheter material often being cited (1,2,3,4,5). Polytetrafluoroethylene (PTE), polyvinylchloride (PVC), polypropylene, and silastic are the materials used most frequently. In addition to their chemical properties, the flexibility and propensity for kinking of these materials have a direct influence on the local reaction (5,6). Highly flexible catheters of shorter length tend to reduce contact with the endothelium and thus decrease the risk of thrombophlebitis (7,8). Other variables that have been evaluated in vascular reactions following catheterization include good surgical preparation of the catheter insertion site (9,10,11) and good dexterity and experience of the manipulator (12); these factors are reported to reduce the risk of thrombophlebitis. On the other hand, serious illness and the injection of irritant products (13) increase the risk of thrombophlebitis.

Commercial PVC tubing is comparable with long-term silastic catheters in its flexibility and resistance to kink. It is now used experimentally as venous and arterial catheters, and it could represent an affordable alternative to commercially available catheters made out of PTE.

The objectives of this study were 1) to compare the thrombogenic properties of a commercial material catheter with those of a customized PVC-like catheter, and 2) to evaluate the effects of external factors on the development of thrombophlebitis in hospitalized dairy cows.

Materials and methods

Both studies were approved by the Institutional animal care committee of the University of Montréal.

Study 1

Twenty purebred, clinically healthy, Friesian Holstein cows, aged from 22 mo to 5 y, were used. During the study, they were kept in a free-stall environment and had access to food and water.

Preparation of the catheter insertion site was standardized as follows: the skin over the external jugular vein was shaved, first with an electric clipper (No. 40 blade) and then with a razor blade, from the angle of the mandible to the thoracic inlet. It was then washed with iodonized soap and water, scrubbed (E-Z Scrub 201; Becton Dickinson, Mississauga, Ontario) for 5 min, and rinsed with 3 alternative passes of alcohol (70%) and 10% povidone iodine solution USP (Iodovet; Rougier, Chambly, Quebec). Local anesthetic, 0.5 mL of lidocaine (Lurocaine, 20 mg/mL; Vetoquinol, Joliette, Quebec), injected at the insertion site, facilitated the insertion of the catheter.

Ten animals were catheterized by using a commercially available polytetrafluoroethylene (PTE) over-the-needle catheter (Angiocath; Becton Dickinson Infusion Therapy System, Sandy, Utah, USA). The PTE catheter had a 14 G external diameter and was 133 mm long. Ten other animals were catheterized with a gas sterilized PVC-like tubing with an internal diameter of 18 G and an external diameter of 14 G (Tygon; Norton Performances Plastics Corp, Akron, Ohio, USA) cut to 150 mm (Figure 1A). This tube was placed IV through a 12 G PTE catheter (Angiocath; Becton Dickinson) just previously inserted and immediately removed following the introduction of the PVC-like tubing. Approximately 130 mm of the PVC tubing were inserted in the jugular vein. The PTE and PVC-like catheters were fixed in place with 2 nylon sutures (Supramid 1; Serag Wiessner, Naila, Germany). An IV extension set (Venotube; Abbott, Sligo, Ireland) with an adaptor (PRN Adaptor; Becton Dickinson) completed the PTE catheter system, while a 5-mm, 18 G PRN adaptor completed the PVC-like catheter system. In both catheter systems, the extension tubes were attached to the skin by a 3rd suture (Figure 1B). All catheters were placed by an experienced manipulator who had previously placed at least 50 catheters.

Figure 1. (A) Material used for study 1; Commercial polytetrafluorethylene catheter versus customized polyvinylchloride-like tubing. (B) Site of insertion and fixation of the catheter.

Figure 1. Continued.

The catheters were kept in place for 120 h; every 12 h, the catheters were flushed with 8 mL of heparinized saline (10 IU/mL). Daily clinical evaluations of each animal including head posture; temperature; appetite; and neck palpation to detect pain, induration, heat, and edema were performed. The catheterized jugular vein was filled by manual pressure to verify patency.

All jugular veins were evaluated echographically (Aloka SSD 118; Aloka, Japan), alter being distended from the mandible to the thoracic inlet, in a transversal and longitudinal cut by using a sectoral 7.5 Mhz probe with acoustic gel (Universal Medical Systems, New York, USA), before catheterization, and before and after removal of the catheter at 120 h. The parameters that were evaluated were the thickness of the vein wall at the insertion site and at the level of the catheter tip; the integrity of the jugular valve; the mobility of the vessel wall, as compressed by the probe; and the presence and description of abnormalities or thrombi (13). At the time of their removal, the catheters were freed and retracted 3 cm, and a sterile scrub preceded complete removal. The last 2 cm of the catheters were removed, using a sterile technique, and then cultured in a broth culture medium both aerobically and anaerobically. All echography evaluations were performed by the same evaluator.

Study 2

Fifty adult dairy cows affected by different diseases and referred to the Veterinary Teaching Hospital (VTH) at the College of Veterinary Medicine at the University of Montreal were used. The reasons for catheterization were IV fluid therapy, antibiotic treatment, or both. To be included in the study, these cows had to be free of clinical and echographical evidence of abnormalities of the external jugular vein. Cows were kept either in tie-stalls with a halter or in box stalls. The same preparation of the incision site and the same clinical and echographical evaluations were used as in study 1. The type of catheter used for all cows was the commercial PTE (Angiocath, 14 G, 133 mm; Becton Dickinson), previously described. Catheters were usually removed after 4 d. All catheters were flushed with 8 mL of heparinized saline (10 IU/mL) every 8 to 12 h, depending on the frequency of the antibiotic administration. The jugular veins from any animal that had to be euthanized for clinical reasons were recovered and dissected to confirm the echographic observations.

The following variables with the potential to produce reactions in the vein were analyzed: the experience of the manipulator (the manipulator was considered to be proficient if he or she had inserted at least 50 catheters; catheterizations were performed first by veterinary students, if they were present at the arrival of the referred animal, or by experienced manipulators, if students were unavailable when the animal was received or catheterized), problems at time of insertion, length of time the catheter was in place, type of fluids (0.9% saline or isotonic salt mixtures, 5% dextrose, antibiotics) administered, body temperature, heart rate, hydration status, active infectious processes, cutaneous reaction at withdrawal time, and suspicion of endotoxemia during treatment. Endotoxemia was suspected if neutropenia with a left shift was observed once on the hemogram. Animals having with a severely inflammatory hemogram will be referred to as endotoxic in the following text to facilitate the reading.

Statistical analysis

Data were analyzed by using a logistic multiple regression model with a backward elimination procedure of the observed variables (Proc. Logistic, version 8.2; Statistical Analysis Systems (SAS), Cary, North Carolina, USA).

Briefly, dependent variables (thrombus presence or absence and vein reaction or not) were analyzed versus independent variables (experience of the manipulator; severe inflammatory hemogram; more or less than 4 d of cannulation; antibiotherapy; saline; dextrose 5%; echographically visible thrombus; vein reaction; body temperature at time of catheterization: < or > 39.2°C; dehydratation: > or < 7%; presence of infectious disease). A multiple regression analysis was used to determine the combination of factors associated with the development of thrombophlebitis. Variables with a P < 0.05 were kept in the final model. Exponentiation of the coefficient parameters in the final logistic regression models yielded odds ratios.

Results

Study 1

In all 20 animals, cardiac and respiratory rates were normal at the time of catheter insertion and withdrawal. Rectal temperature remained within the normal range throughout the study. All jugular veins appeared macroscopically normal and remained patent during the study period. In each catheter group, 1 animal pulled out the catheter: the animal in the PVC-like group after 2 d and the animal in the PTE group after 4 d. Furthermore, a PTE catheter was severely damaged at its base at the insertion site.

All other catheters were patent until removal at 120 h. A thrombus was found at the insertion site, 4 to 10 mm in size in 3 animals from the PVC-like group, and 10 to 20 mm in size in 5 animals from the PTE group (Figure 2A, 2B). None of the thrombi completely obstructed the blood flow. All the PTE catheters were kinked and 30% leaked at the kink. All PVC-like catheters were intact. No bacterial growth, either aerobic or anaerobic, was found in either group (Table 1).

Figure 2. Macroscopic (A) and echographic (B) appearance of a thrombus localized at the catheter insertion site. Also note the presence of normal valves (A).

Figure 2. Continued.

Table 1.

Study 2

Table 2 presents the reasons for referral of the 50 hospitalized cows, and Table 3 presents the variables studied and their distribution before analysis in the logistic regression model. Thirty-six cows were catheterized for 4 d or less (one cow 3 d) and 14 cows for more than 4 d, including 1 cow for 7 d.

Table 2.

Table 3.

Clinical signs of severe phlebitis, such as vein induration, pain on palpation, abnormal head posture associated with loss of appetite, or increased rectal temperature, were not observed in the majority of the 50 animals. One animal developed a small abscess and 2 others developed a hematoma at the site of insertion. Only 1 animal had pain on jugular palpation. All jugular filling times were normal, except in 1 animal, which had thrombophlebitis and thus a slowed refill time.

Echographically, 12 of the 50 animals developed a single thrombus and 1 other had 2. Nine thrombi were located 1 to 7 cm distal to the tip of the catheter, always at the first jugular vein valve downstream from the catheter (Figures 3A, 3B). Two thrombi were directly at the tip of the catheter. The other 3 thrombi were at the sight of insertion. On 3 other animals, the wall of the jugular vein was thickened at the site of insertion for a few centimetres, without thrombus formation; 2 valves were also thickened, 1 had been perforated by the catheter, and the other was 5 cm distal to the tip of the catheter (Table 4). Jugular dissection of the animals euthanized during, or after, the study correlated with the echographic findings (Figures 2, 3).

Figure 3. Macroscopic (A) and echographic (B) appearance of a thrombus localized at the first valve downstream from the catheter tip.

Figure 3. Continued.

Table 4.

Analysis of the data revealed only 2 significant (P > 0.05) factors associated with a reaction in the vein: the presence of a severe inflammatory hemogram or endotoxemia was associated with thrombus formation, and the jugular vein reaction was inversely related to the manipulator's experience (Figures 4A, 4B). In endoxotic cows, the development of a jugular thrombus was 7.8 times (P = 0.0297, confidence interval of Wald 1.22 to 49.39, 95%) more likely than in a nonendotoxic cow. If the manipulator was inexperienced, the cow was 6 times (P = 0.0077; confidence interval of Wald 0.045 to 0.622, 95%) more likely to develop a jugular reaction. No significant relation (P > 0.05) was observed between the presence of endotoxemia and the length of time the catheter remained in the vein or with the other evaluated variables.

Figure 4. Effect of endotoxemia and manipulator experience on the presence of thrombus formation (A) and vein reaction (B), respectively.

Figure 4. Continued.

Discussion

In study 1, the thrombi (8 out of 20 animals) were all found at the insertion site. Similar observations were made by Pusterla et al (9), who associated the traumatic breach in endothelium, followed by activation of the coagulation cascade, coupled with possible infection at the insertion site, with the initiation of the thrombus formation. The negative bacterial cultures indicate that infection did not seem to be a related factor in the present study. On the other hand, many PTE catheters were broken and others leaked, which caused mechanical irritation to the intima near, or at, the insertion site. The animals used in study 1 were young and vigorous, kept in free stall, and they had to pass their heads through metallic gates to gain access to feed, all factors that might increase catheter movement and thus trauma to the vascular wall. These elements were absent in study 2, since the cows were kept individually in a free stall without head gates or tied with a halter, and no gates were present to access feed. This could explain why only 3 out of 50 animals developed thrombi at the insertion site, of which 2 had heavily damaged catheters and in 1 there were problems at time of insertion.

Many studies have tried to elucidate the question about the type of material used for catheterization and its effect on thrombus formation. Welch et al (3) and Néjad et al (2) reported marked thrombogenic reactions to polyethylene catheters and minor reactions to silastic. Jones et al (1) reported a higher frequency of thrombosis when PVC catheters were used, but no difference between propylene and teflon materials. Spurlock et al (7) reported that the highest reaction was obtained with teflon, followed by polyurethane, with silastic causing only mild reaction. These studies agreed that the physicochemical properties of a catheter, its length, and its compliancy were important in determining the risk of thrombus formation.

In study 1, all of the rigid-type catheters were damaged at time of removal versus none of the PVC-like ones. The possibility of using a safe, low cost, and flexible catheter under field conditions is, therefore, very interesting. Since these catheters are more resistant to damage, they may be left in place with less hazard and can be changed easily on a regular basis by using a 12 g needle. The PVC-like material used in study 1 gave satisfactory results with regards to durability, low thrombogenicity, and safety, while keeping an adequate function at low cost (less than $1.00). This type of material might be useful for long-term IV drug administration, but an internal diameter of 18 G is too narrow for fluid administration.

The most common complication associated with catheterization is thrombophlebitis. Thrombophlebitis is defined as thrombus formation associated with inflammation of the vein. Since the presence of a thrombus by itself creates inflammation, it is unusual to observe a thrombus without phlebitis (14). A cavitated thrombus is usually the sign of a septic thrombus (15). The absence of thrombus cavitation on echography in both studies, as well as at necropsy, plus the negative results on culture, allowed the presumption that all thrombi formed were aseptic in both studies. Since many cavitated thrombi have been reported to develop at the insertion site (15) and preparation of the injection site is an important factor in the development of thrombophlebitis (16), the absence of cavited thrombi in the present study indicates adequate insertion site preparation. Septicemia and pulmonary embolism associated with anorexia, fever, or poor general condition are possible sequelae to thrombophlebitis (9), but they could not be related clinically to the venous reaction in the present study.

The severity of disease process presenting conditions seems to be an important factor to the development of jugular complications. In the present study, the animals with severe inflammatory hemograms were 7.8 times more likely to develop thrombophlebitis than were animals with non-severely inflammatory hemograms. A severely inflammatory hemogram may reflect endotoxemia. Endotoxins have been shown to interfere with clotting (14,17) by affecting endothelial cells and exposing subendothelial collagen (14), by decreasing thrombomodulin production and heparin activity of the endothelial cells (14), by increasing the production of the inhibitor of plasminogen activator, and by decreasing the production of protein C and plasminogen activator. The consequence is an excessive production of fibrin and decreased fibrinolysis, with a net increased risk of thrombogenesis (14,18). Endotoxins also have a direct effect on platelets and the coagulation cascade (17). Hemostatic imbalance created by circulating endotoxins increases the risk of thrombosis, thus enhancing the risk of complications following catheterization and indicating the importance of reducing trauma to a minimum when inserting the catheter. A compounding effect of endotoxemia is the implementation of a more aggressive treatment regimen, involving more frequent injections, an increased number of drugs used, more intensive fluid therapy, and, consequently, an increased manipulation of the catheter, which in itself increases the risk of complications.

The results of this study clearly demonstrate that jugular vein reactions are inversely related to the experience of the manipulator. In humans, a reduced frequency of thrombosis has been associated with the use of experienced personnel (12,16), resulting in fewer unsuccessful attempts to insert the catheter, since multiple insertions lead to more endothelial damage and perivascular hematoma formation. In the horse, a relationship between complications associated with catheterization and the manipulator's experience and the number of attempts before successful insertion has been speculated (19).

Rubbing contact between the endothelium and the tip of the catheter has often been implicated as a potential thrombus initiator (2,13). In these 2 studies combined, 20 thrombi were found. Only 2 were observed at the level of the catheter tip. Eighty-six percent (60/70) of the catheters used were rigid (PTE), indicating that catheter rigidity and rubbing against the intima might play only a minor role in thrombus formation in cattle (13). Conversely, the first valve downstream of the catheter tip was incriminated in the majority of thrombus formations. Pusterla et al (13) reported that the valve was a predilection site for thrombus formation with or without complete obstruction to blood flow. In addition to the increased turbulence and predilection of the site for thrombus attachment (13,19), it has been postulated that an increased amount of coagulation factors in the environment of the valve might be implicated in thrombus formation (3,16). In the present study, some valves were in contact with catheter tip and others were as much as 7 cm away from the catheter tip. Injection of irritant drugs and blood turbulence created by injection, associated with the previously mentioned main predisposing factors, may explain the thrombus formation at that site. Contrary to other studies (13), no significant relation has been established between type of fluids infused, injected, or both, and thrombus formation.

In conclusion, this study shows that PVC-like tubing might be less thrombogenic than a rigid-type catheter might be in freely moving animals, and that for a rigid-type catheter, the experience of the manipulator and the severity of the illness are 2 predisposing risk factors for the development of thrombophlebitis in dairy cattle.