Surveillance of Infections Associated With Intravenous Catheters in Dogs and Cats in an Intensive Care Unit

Introduction

Infections associated with intravenous (IV) catheters are one of the top two causes of nosocomial infections in hospitalized humans and animals, and they include catheter colonization, skin exit-site infection, and catheter-associated bloodstream infections.^1–6 Reported routes of infection include skin flora migration down the exterior wall of the catheter to colonize the tip, contamination of the catheter hub with bacteria that migrate down the lumen of the catheter, seeding of the catheter with bacteria from distant sites in the body, and contamination of the infusate.^1,4,5,7–13 Contamination of the catheter by skin flora is reported to be the most common route, followed by contamination of the catheter hub, with colonization from distant sites of infection and infusate contamination being reported as rare.^1,8,11,14

Catheter-associated infections in humans are a frequently reported cause of increased patient morbidity and mortality, and increased financial cost to medical institutions.^{1,5–7,15–17} Catheter-associated infections have been researched and well documented in humans and have led to the development of protocols aimed at decreasing their risk and occurence.^{1,2,4–11,14–16,18–24} Similar prospective studies in small animals are lacking. The purposes of this study were to prospectively assess positive catheter-tip cultures associated with IV catheters in cats and dogs hospitalized in an intensive care unit (ICU) and to identify potential risk factors that are associated with increased positive culture rates in these animals.

Materials and Methods

Any dog or cat admitted to the ICU from November 2002 through November 2003 (phase I) and July 2004 through November 2004 (phase II) in which an IV catheter was placed and maintained for a minimum of 48 hours was eligible for inclusion in the study. Cases were entered into the study if proper protocol was followed for catheter insertion and removal and if the catheter was placed in the ICU. The study was performed in two phases in an attempt to increase the sample size. Catheters used in this study included 16- and 18-gauge, 28-cm, polyvinylchloride central venous catheters^a inserted into the jugular or saphenous vein; and 22- and 20-gauge, 1-inch or 18-gauge, 1.5-inch Teflon catheters^b inserted into the cephalic or saphenous vein. Peripherally placed catheters were used for administration of crystalloid and colloid fluids, blood products, and IV medications. Centrally placed catheters were used for the same purposes, as well as for total parental and partial parental nutrition administration, dextrose infusions, blood sampling, and central venous pressure monitoring.

A strict protocol for catheter placement was followed. A large region of hair around the vein was clipped to prevent potential contamination of the insertion site with hair. The skin over the vein was alternately scrubbed with 2% chlorhexidine^c surgical scrub and wiped clean with 70% isopropyl alcohol^d three times. The person inserting the catheter either wore gloves or washed hands with antibacterial soap for 2 minutes prior to insertion. After placement of central venous catheters, bacitracin-neomycin-polymyxin antibiotic ointment^e was applied over the insertion site and covered with a gauze sponge. The needle guard was secured to the animal with adhesive tape. A bandage was applied to the area using rolled gauze and a self-adherent wrap. After peripheral catheters were inserted, they were simply secured to the limb with adhesive tape.

The IV catheters were replaced or removed when they were no longer needed or became nonfunctional, if evidence of tenderness or inflammation involving the vein or surrounding tissues occurred, or if the animal developed a fever. The catheter site was assessed daily for signs of swelling, tenderness, or erythema, and the vein was palpated for signs of phlebitis. Bandage changes were performed if the bandage became loose, wet, or soiled. During crystalloid and colloid administration, IV lines were changed each time a new bag of fluid was started and if the line became contaminated or soiled. Total parenteral nutrition (TPN) and partial parenteral nutrition (PPN) bags and lines were replaced every 24 hours. Blood products were administered through a separate administration set that was discarded after completion of the transfusion.

A strict protocol was also followed for removal and culture of the catheters. The bandage was removed, and the catheter insertion site was scrubbed with 2% chlorhexidine scrub, wiped with 70% isopropyl alcohol, and allowed to air dry. When the catheter was removed, contact with the animal was avoided, and the distal 1 to 2 cm was cut with sterile scissors and placed into a sterile, red-top blood collection tube. Sterile 0.9% sodium chloride (1 mL) was added to the tube to prevent drying of the catheter tip. Catheters were submitted to a commercial laboratory for aerobic and anaerobic bacterial culture and were kept at room temperature until processing, which was performed within 16 hours of submission. The fluid sample from the red-top tube was vortexed and then cultured on blood agar^f and MacConkey agar^g for aerobic bacteria, and on Brucella-laked blood with kanamycin and vancomycin agar^h for anaerobic bacteria. The catheter tip was rolled across all three agar plates prior to inoculation of the agar with the fluid sample. A culture was considered positive if there was any growth on the plating media within 72 hours for aerobic cultures and within 7 days for anaerobic cultures. A three-quadrant streak for bacterial isolation was used, and growth was classified as light (growth in quadrant one only), moderate (growth in quadrants one and two), or abundant (growth in all three quadrants). A specific colony count was reported when individual colonies could be enumerated. All positive cultures were further evaluated for identification of bacterial species.

Upon removal of the IV catheter, the insertion site and surrounding area were evaluated for signs of inflammation, tenderness, or phlebitis. The medical record was reviewed, and a form was completed to document the risk factors that were present for each animal. Positive catheter-tip culture rate and type of bacteria cultured were also evaluated. Risk factors that have been reported with catheter-related infections in humans were also assessed for their potential association with positive cultures in this study.^{1–8,10,13–15,17–19,23,25–32} Risk factors assessed included blood sampling from the catheter, infusate type (e.g., TPN, PPN, dextrose, blood products, and fluids designed for oncotic support), duration of catheterization (i.e., ≤72 hours versus >72 hours), catheter type, catheter location, and catheter complications (e.g., swelling, erythema, phlebitis/thrombosis, wet bandages, and kinks). Development of fever and outcomes (e.g., discharged, died) were also evaluated.

Univariate statistical analysis for association of risk factors and positive catheter-tip culture were performed with the chi-square distribution or Fisher’s exact test, depending on sample size.³³ A P value of <0.05 was considered statistically significant. A multivariate statistical analysisⁱ was then applied to the data to determine if an interaction existed among two or more possible risk factors. The Fisher’s exact test was used when an expected value was <1 or if <20% of the expected values within a multivariate analysis were >5. When the chi-square test was used, the Yates’ continuity correction was applied.³³ This correction is designed to make the approximate results from a chi-square test more accurate when applied to small samples. A two-sided P value was considered, and significance was determined if P was <0.05.

Results

A total of 101 central and 50 peripheral catheters were submitted for culture from 113 dogs and 38 cats. Thirty-seven samples were positive, for a total positive culture rate of 24.5% (37/151). Of the positive cultures, 81.1% were from dogs (n=30) and 18.9% were from cats (n=7) [Table 1]. The rate of positive catheter cultures between dogs and cats was not statistically significant. Four animals (three dogs, one cat) had multiple bacteria isolated, including two bacteria from two dogs and one cat and three bacteria from one dog. Ten different bacterial isolates were cultured. Enterobacter spp. was the most common isolate, accounting for 46.0% (17/37) of the total positive cultures [Table 2].

Risk factors assessed for association with a positive catheter-tip culture included blood sampling through the catheter (20/86 positive cultures, P=0.37), TPN administration (8/25 positive cultures, P=0.34), blood product administration (5/12 positive cultures, P=0.14), PPN administration (1/1 positive culture, P=0.25), dextrose infusion (2/6 positive cultures, P=0.46), hetastarch infusion (2/13 positive cultures, P=0.88), catheter duration of ≤72 hours (12/58 positive cultures, P=0.35), and catheter duration of >72 hours (25/94 positive cultures, P=0.35). Other risk factors assessed were catheter type (including venocath catheters [25/101 positive cultures, P=0.35] and terumo catheters [12/50 positive cultures, P=0.37]), location of the catheter, and type of catheter used at each location (including cephalic catheters [12/48 positive cultures, P=0.37], jugular catheters [19/84 positive cultures, P=0.36], saphenous terumo catheters [0/3 positive cultures, P=1.0], saphenous venocath catheters [6/17 positive cultures, P=0.21], and total saphenous catheters [6/20 positive cultures, P=0.36]). Also assessed were catheter complications, including swelling (3/10 positive cultures, P=0.46), phlebitis or thrombosis (4/14 positive cultures, P=0.46), wet bandages (1/2 positive cultures, P=0.43), kinking of the catheter (1/2 positive cultures, P=0.43), erythema at the catheter site (1/1 positive culture, P=0.25), fever (7/16 positive cultures, P=0.06), and combined catheter complications (10/29 positive cultures, P=0.32). Case outcomes, including animals that were discharged from the hospital (33/136 positive cultures, P=0.74) and those that died (4/15 positive cultures, P=0.53), were also analyzed. None of the risk factors evaluated had a statistically significant association with positive catheter-tip culture.

Discussion

In this study, the overall positive culture rate was 24.5%, compared to 5% to 47% in humans for all types of IV catheters and 22.4% to 52.2% for central venous catheters.^4,12 Prior reports have shown catheter-associated infection rates of 15.4% to 48.9% with peripheral IV catheters in dogs, 26.0% with jugular catheters in dogs and cats, and 10.7% with both catheter types in dogs and cats.^25,34,35 The last study used >5 colony-forming units (CFU) to define a positive culture, whereas the former studies considered any bacterial growth as a positive result, which may explain the lower positive culture rate observed in the later study.^25,34,35 The positive IV catheter-tip culture rate reported in this study was within the range of positive culture rates reported in previous human and animal studies.^4,12,25,35

Maki et al. described a semiquantitative culture method for IV catheters that interpreted positive cultures with <15 CFU as being indicative of catheter colonization rather than septicemia.¹⁴ This study has been criticized for not taking colonization of the catheter lumen into account, as the exterior of the catheter was rolled over the culture media during culturing.^4,6,11,16 Quantitative culture has also been described, where the catheter tip is vortexed or sonicated, allowing for removal of bacteria from the lumen of the catheter as well as from the exterior surface.^1,4,6,7,11 The study reported here used a combination of these techniques. The catheter tip was submitted in 1 mL of sterile saline, which was vortexed. The saline was cultured, and the catheter tip was also rolled across the culture media. This combined technique may have increased the chance of a positive culture result. Any bacterial growth was considered a positive result in an attempt to increase the sensitivity of the study. Colony counts were done on those cultures for which individual colonies could be enumerated. If cultures with colony counts <15 CFU were excluded from the study, the positive catheter culture rate would decrease to 16.6% (25/151). If the conclusions from the Maki et al. study are taken into consideration, this may indicate that the catheters with <15 CFU (7.9% of total positive cultures) were contaminated rather than associated with infection.

Anaerobic cultures were also performed on all samples, and only one culture was positive, with moderate growth of Bacteroides spp. in a cat. Positive anaerobic culture results may have been artificially low, as the samples were not submitted under strict anaerobic conditions.

Studies in humans have reported skin flora as being the most common bacterial isolates, with coagulase-negative Staphylococcus (epidermidis) and Staphylococcus aureus as the primary isolates.^{1,2,4–7,9,11,26,36} Skin contamination during catheter insertion and use, and migration of bacteria from the skin down the exterior of the catheter tip have been implicated as sources.^{1,2,4–7,9,11,14,26,36} Although skin flora are the most common isolates, the frequency of isolation of gram-negative and enteric organisms increases in ICU patients that are severely ill.^7,37 In the study reported here, the three most common isolates were not normal skin flora and included Enterobacter spp., Escherichia coli, and Pseudomonas spp. Hub contamination and migration of bacteria down the lumen of the catheter was of concern in these animals. Contamination of the infusate could not be ruled out but seemed less likely, as multiple types of infusate were used from various sources. Culture of the infusate at the time of catheter removal would have provided more information regarding the source of contamination. Distant sites of infection may also have accounted for some catheter-tip colonization in animals of this report, but such colonization is rare, and distant-site infections were ruled out by a thorough diagnostic workup.

A high rate of isolation of Enterobacter spp. occurred in this study, with 16/17 of these cultures obtained in the first phase of the study (2002–2003) and only 1/17 obtained in the second phase (2004). The presence of Enterobacter spp. was suspected to be a nosocomial infection with possible transmission of bacteria from human hands or contamination of fomites used in catheter insertion or maintenance. Soiling of the catheter with urine, feces, food, or saliva was also possible but less likely, given the marked difference in positive catheter-tip culture rates in phase I compared to phase II of the study. The cages housing the animals, cleaning methods, and protocols were the same between the two phases of the study. Differences between the two study phases included staff turnover and the batches of materials used for preparation and placement of IV catheters. A marked decrease in positive Enterobacter spp. cultures occurred in the second phase of the study, which might indicate a point source of infection was eliminated during the 7-month lag phase between the two study phases.

Multiple risk factors associated with IV catheters have been identified in humans, including duration of catheterization, catheter location, blood sampling from the catheter, infusate type, phlebitis, and catheter type.^{1–8,10,13–15,17–19,23,25–32} Duration of catheterization >72 hours has been reported to increase the risk of phlebitis and subsequent catheter-related infections in humans.^{2–4,8–10,14,17,19,30} In contrast, several animal and human studies have suggested that routine replacement of catheters every 72 hours may not decrease the risk of catheter-associated infections.^{1,3–8,10,17,34,35} The current study shows no association of positive cultures with duration of catheterization or with catheter complications.

An increased risk of infection from central venous catheters placed in the jugular vein has been reported in humans when compared to catheters in the subclavian and femoral veins.^1,4,6,8,20 In dogs and cats, saphenous catheters may be at greater risk for soiling with urine and feces; cephalic catheters are more accessible for licking, chewing, or soiling with food; and jugular catheters are in closer proximity to oral and respiratory secretions. Despite these concerns, no association between catheter location and positive cultures is seen in the study reported here. In humans, conflicting reports exist as to whether the risk of infection is greater with peripheral catheters versus central venous catheters.^{2,4,5,8,18,29,30} The present study shows no statistically significant association between a positive culture and either catheter type.

Blood sampling from central venous catheters has also been reported to increase the risk of catheter-associated infections in humans, possibly from manipulation of the catheter and repeated penetration of the hub, with increased risk of contamination.^3,8,29,30 Most (85.2%, 86/101) of the central venous catheters in the current study were used for blood sampling; however, no association was found between blood sampling and positive cultures in this study. Infusate type has also been reported to be a risk factor in humans. Hypertonic solutions such as dextrose increase the risk of phlebitis, which is thought to predispose to catheter-associated infections.^3,10,30,31 Lipid emulsions and blood products also have a propensity to promote bacterial growth.^{7,10,13,23,26,29–31} No statistical association was found between positive cultures and any individual type of infusate when administered alone or in combination.

Some prior reports have indicated higher rates of catheter-associated infections in the presence of phlebitis.^3,10,17,25 Other studies, including the present one, have found no such association.^4,7,28 In the animals of this study, it is interesting to note that phlebitis occurred in some animals that had negative cultures (n=10); some animals had a positive culture with no evidence of phlebitis (n=33); and only four animals had both a positive catheter-tip culture and phlebitis.

Catheter-related infections have been associated with increased morbidity and mortality in humans, but their clinical significance has been hard to characterize, as it is often difficult to determine whether the infection or the underlying disease is the primary cause of morbidity.^{1,5–7,15,16} In the present study, there was no association between case outcomes and the culture results. Fever can result from underlying infection, but it was not associated with positive cultures in the present study. Bloodstream infections associated with IV catheters cannot be determined without concurrent positive blood cultures and isolation of the same organism from both sites, as a positive catheter-tip culture alone does not necessarily indicate the presence of infection or bacteremia.^{1,5,6,9,25,26,34} A positive culture only indicates that the IV catheter is contaminated, which might predispose to bacteremia.^26,27 One study that assessed long-term, indwelling central venous catheters reported no clinical signs in eight research dogs that had positive catheter-tip cultures and concurrent positive blood or tissue cultures.³⁸ Bacteremia was not ruled out in the study reported here, because blood cultures were not performed.

Skin contamination upon catheter removal seemed an unlikely source of positive bacterial cultures in this study population, in that the skin around the catheter site was prepared with a surgical scrub prior to catheter removal and most of the isolates were not normal skin flora. Contamination of catheters without infection was likely in some of the animals, as local and systemic signs of infection were lacking and some of the catheters grew <15 CFU (n=12), which has previously been shown to be consistent with bacterial contamination without septicemia.¹⁴ Because 35.2% (6/17) of the positive Enterobacter spp. cultures yielded <15 CFU, contamination of these catheters could not be ruled out. Semiquantitative or quantitative culture methods, including blood cultures, would have provided more specific information as to whether the catheter tips that grew positive bacterial cultures were contaminated rather than associated with infection and bacteremia.

Conclusion

In the current study, no statistically significant associations were found between a positive catheter-tip culture and risk factors that have previously been reported to increase the risk of IV catheter-associated infections in humans. Risk factors assessed included blood sampling from the catheter, infusate type administered, duration of catheterization, catheter type and location, catheter complications, and case outcomes. A nosocomial source of Enterobacter spp. was suspected, given the high culture rate of this organism in phase I of the study. In future studies, concurrent catheter-tip and blood cultures would provide more information regarding bacterial contamination of the catheter tips versus true catheter-associated infection.

Acknowledgment

The authors thank Kevin Hahn for his help with the statistical analysis and review.