Peripherally inserted central venous catheters are frequently used to provide prolonged intravenous access for administration of drugs and fluids, sampling of blood, or hyperalimentation.1, 2 Compared with other tunnelled central venous catheters, catheters inserted peripherally are easier to place and remove.3 In previous studies, use of such catheters was associated with lower rates of complication compared with other tunnelled central venous catheters.4–11 Other potential advantages include fewer venous punctures, and decreasing pain and venous irritation from therapeutic delivery of drugs. Furthermore, they allowed for early hospital discharge.12
Neonates with critical congenital cardiac disease are a special population. They are often critically ill, and have need for safe and prolonged central intravenous access to provide drug therapy and allow sampling of blood, since they have poor tolerance to the stress of repeated venepuncture. The objective of our study was to evaluate the use of such catheters inserted peripherally to provide central venous access in this unique population, as well as to study features such as the duration of use, reasons for removal, and complications.
Methods
Population studied
This study presents our experience of use of peripherally inserted central venous catheters in neonates with critical congenital cardiac disease who were admitted to the Intensive Care Unit at Texas Children's Hospital from August, 2002, to August, 2004. The catheters were used for multiple purposes, including intravenous delivery of fluids, administration of medications such as inotropic agents, total parenteral nutrition, and antibiotics, sampling of blood, and monitoring of central venous pressure.
After obtaining institutional review board approval, we retrospectively reviewed the medical records of all the neonates with critical congenital cardiac disease who had undergone peripheral insertion of central venous catheters during the period of study. Data recorded included: birth weight, gestational age, weight and age at catheter insertion, diagnosis, characteristics of the catheter and duration of its use, reasons for removal, and catheter-related complications.
Method of placement of the catheters
Referrals for insertion of the catheters occurred at the discretion of the primary physician, and were made directly to the interventional radiologist, who inserted the catheter. Because of ongoing cardiac disease, the patients were transported by the cardiovascular anaesthesiologist, who then managed sedation and anaesthesia.
Preparation and insertion were standardized by the Interventional Radiologists. The majority of catheters were inserted in the arms. The vein punctured was identified with high-resolution ultrasound (Phillips HDI 5000). After a suitable vein for insertion was identified, the area of the skin at the proposed site was prepared and draped. The interventional radio-logist inserting the catheter wore a cap, mask, sterile gown, and sterile gloves. A 21Gauge vascular access needle was advanced under direct ultrasonic visualization into the vein. A tunnel of at least 2 centimetres was created between the site of puncturing the skin and the site of venous entry. Once the needle was in place, a guide wire was advanced into the superior caval vein. The length of the catheter was determined by measuring the guide wire. The catheters, mostly COOK silicone and later Deltec polyurethane, were then placed using an “over the wire” technique. Fluoroscopy was very helpful in maneuvering the guide wire, which often required manipulation to reach the superior caval vein. The length was carefully checked by moving the arm or leg of the patient, and adjusted as needed to rest at the cavo-atrial junction13 confirmed by fluoroscopy. If difficulties were encountered, injections of non-ionic contrast were made to assess venous anatomy. Occasionally, the femoral veins were used. The femoral catheter was tunnelled from a site of skin puncture above the knee to a venous entry point in the femoral or saphenous vein in the groin. Afterwards, the catheter was sutured in place and a sterile dressing applied. Care of the catheter was administered by the nurses caring for the patient. The dressing was changed every week, or earlier if soiled. Catheters were accessed continuously or intermittently. For peripherally inserted central venous catheters accessed continuously, concomitant 1 unit per millilitre heparin-containing solution was used at a rate of 1 millilitre per hour to maintain patency. When the catheters were accessed intermittently, or having drawn blood, they were flushed with heparin-containing saline solution and locked after each use.
End points
All catheters were monitored clinically for complications. Complications were recorded including occlusion, accidental dislodgement, mechanical malfunction due to damage to the catheter or unplanned withdraw, thrombosis, and catheter-related infection. All complications necessitated removal of the catheter. Dislodgement was defined as displacement of the catheter tip with extravasations of the fluid outside the vein. Thrombosis was defined as swelling of the arms, neck or face in infants with catheters placed in the arm, or swelling of the leg in infants with femoral venous catheters. If suspected, thrombus was confirmed by ultrasonography and/or venography. The following definitions were used for infectious complications:
- phlebitis: inflammation tracking along the course of vein from the site of insertion, with or without a palpable venous cord
- exit-site infection: inflammation and purulent discharge were noted at the site of insertion
- catheter-related sepsis: a positive culture from blood or catheter in the presence of an indwelling catheter
The incidence rate of catheter-related infection was calculated as the number of first culture positive blood stream infections divided by the number of patient-catheter days at risk, in other words, the number of days from the day of insertion until the day of the first positive blood culture or until the catheter was removed for some other reason.
Statistical analysis
The characteristics of the patients, and factors related to the use of the catheters, such as age, sex, body weight, prematurity, palliative or corrective cardiac surgery, catheter size and lumen, tip position, duration of use, and hospital stay, were examined univariately to identify the risk factors associated with overall complications, as well as two commonly seen complications, namely catheter-related infection and occlusion, at the 0.10 α level. In the multivariate analysis, a multiple logistic regression model was used. Factors included in the multivariate analysis were those identified by univariate analysis at a significance level of 0.10 as indicated. Odds ratios and 95 per cent confidence interval were computed for the significant factors. The analysis focusing on catheter-related infection was performed after restricting data to the initial insertion in order to avoid confounding by multiple successive insertions.
Data are expressed as mean plus or minus standard error of the mean, or median if the data were skewed. All p values were two-tailed, and a p value less than 0.05 was considered significant.
Results
A total of 124 catheters were inserted for intravenous access in 115 neonates with critical congenital cardiac disease. At the time of placement, the patients had a mean age of 10 plus or minus 0.7 days and body weight of 3.1 plus or minus 0.06 kilograms. Median hospital stay was 28 days, ranged from 5 to 349 days. Demographic data is presented in Table 1. Most frequent diagnoses were hypoplastic left heart syndrome in 24%, followed by pulmonary atresia in 17%, and transposition in 13%. During their hospitalization, 79% of the patients underwent cardiac surgery. In two-thirds of the patients, the catheters were placed at a median of 3 days before cardiac surgery or catheterization. One-quarter had other central venous catheters simultaneously, either in the umbilical or the femoral vein. None of the patients had direct right atrial or jugular venous catheters inserted at the time of surgery.
Table 1. Patient demographic data.
Characteristics of the catheters are listed in Table 2. Most frequently, we used 3 French single lumen catheters. The median basilic vein was most commonly accessed. The tip of the catheter was located at the cavo-atrial junction in 94 per cent of insertions. The catheters were in place for 22.3 plus or minus 1.5 days, with a total of 2769 catheter-days. Reasons for removal are shown in Figure 1. Therapy was completed in 76.6% of insertions. The incidence of occlusion, dislodgement, and thrombus was 4.0%, 2.4%, and 1.6% respectively. In 7.3% of cases, the catheters were removed after death of the patient without complications. Neither phlebitis nor infection at the exit-site was noted. The incidence of catheter-related sepsis was 3.6 per 1000 catheter-days, with a median onset of 37 days, and a range from 11 to 137 days after placement. The organisms cultured from blood or catheters were coagulase negative staphylococcus in 6 patients, enterobacter cloacae in 2 patients, pseudomonas aureginosa in 1 patient, and serratia marcescens in 1 patient.
Table 2. Catheter characteristics.
Figure 1. Reasons for removal of peripherally inserted central venous catheters in neonates with critical congenital cardiac disease.
Two or more attempts at insertion of the catheter was required in 23 patients. The rate of complications among these patients was 17.4%, which was not significantly different from 15.8% in those who had only one attempt (chi-square: 0.017, p value equal to 0.895).
No complications occurred during placement or removal of the catheters, and no deaths were directly attributed to their use. One neonate developed a right pleural effusion of moderate size on the sixth day after placement of a 3-French catheter. Fluid drained from the pleural space was similar to the intravenous fluid administered through the catheter, which was then removed. The pleural effusion did not recur, and the neonate recovered. Another patient had major thrombus and developed superior caval venous syndrome on the seventh day after placement and underwent thrombectomy. The symptoms were relieved after operation, and the patient was eventually discharged without further documented complications.
Univariate analysis of the demographic and factors related to catheter use revealed that a prolonged hospital stay was a risk factor for complications. Risk factors for occlusion of the catheters in univariate analysis included lower body weight and a prolonged hospital stay. Longer duration of use was of statistical significance to increase the risk of catheter-related sepsis. In multivariate analysis, a prolonged hospital stay was the factor associated with an increased risk for overall complications, with odds ratio of 1.05, and 95% confidence intervals from 1.03 to 1.08 (p value equal to 0.0001), and also with occlusion, the odds ratio being 1.03, with 95% confidence intervals from 1.01 to 1.06 (p value equal to 0.02). Patients with lower body weight were more likely to occlude their catheters, with odds ratio of 0.06, and 95% confidence intervals from 0.01 to 0.71 (p value equal to 0.03). Longer duration of use was independently associated with catheter-related sepsis, the odds ratio being 1.07, and 95% confidence intervals from 1.02 to 1.13 (p value equal to 0.01). The incremental daily risk of catheter-related sepsis was virtually constant throughout the entire duration of use, with a mean rate of 6% per day.
Discussion
Achieving and maintaining vascular access is a challenge in neonates with critical congenital cardiac disease. In our study, therapy was completed in over three-quarters of insertions, with the average duration of use being 22.3 days after placement. The incidence rate of occlusion, dislodgement, and thrombus was 4.0%, 2.4%, and 1.6% respectively. The infection rate was 3.6 per 1000 catheter-days, with a median onset at 37 days after placement. Catheters inserted peripherally, therefore, were proved to provide a safe and reliable alternative to other central venous catheters for prolonged intravenous therapy in neonates with critical congenital cardiac disease.
The rate of completion of treatment that we achieved was higher than previously reported in either adults or children.7, 8, 14 The image guidance of ultrasound and fluoroscopy routinely used for all vascular access by the skilled interventional radiologist in our study contributed to the shorter time needed for insertion, and fewer attempts. It has been previously reported that the risk of catheter-related thrombosis and infection increases with each attempt at insertion,15 as a result of rupture of the endothelium and interruption of the laminar blood flow by the local trauma of insertion, as well as the physical presence of the catheter itself. This leads to exposure of the blood to the venous intima, which, in turn, can lead to local activation of the coagulation cascade. The process of formation of thrombus can be enhanced by leakage and production of haematomas around the vein. The risk of infection is increased dramatically by such thrombus formation. The thrombus probably serves as culture medium for bacteria, allowing them to multiply more rapidly and easily. Compared with other studies,7, 8, 14 the incidence rate of catheter related infections per 1000 catheter-days was the lowest in this study. Furthermore, infection occurred later, with a median onset of 37 days after placement. Our data show that the risk of catheter-related infection was closely correlated with the duration of use.
Thrombosis associated with the placement of such catheters is thought to be primarily due to venous puncture inducing endothelial injury and phlebitis. Increased rates of phlebitis have been reported with increasing catheter size, 3 French catheters are more likely to be completely occlusive to the vein than 2 French catheters, and thus would be more prone to stasis and thrombosis. Racadio et al.,16 however, found that there was no significant difference in complication rates between use of 2-French and 3-French catheters in children. Recently two studies17, 18 determined the incidence of venous thrombosis in patients with peripherally inserted central venous catheters by using venography, and also revealed that there was no significant correlation between the occurrences of venous thrombosis and the size of the catheter. The majority of our catheters were single lumen and 3-French. The 3-French catheter is the smallest catheter that can meet the need for drawing of blood, which was one of the primary goals for our use of peripherally inserted central venous catheters in neonates with critical congenital cardiac disease. Moreover, Reed et al.19 demonstrated that the risk of infectious complications associated with peripherally inserted central venous catheters was slightly higher when more infusion ports were present. This may relate to a more frequent and/or intense use of the peripherally inserted central venous catheter. All these findings may account for the most often and safe use of single lumen 3-French catheters in our study.
Ideally, the catheter tip should be located just outside the heart and parallel with the long axis of the vein in which it is located.13, 16 The carina can be used as a radiological landmark,20 located approximately halfway along the superior caval vein and marking the beginning of the pericardial sac. There are higher risks of thrombosis if the catheter tip is located high in the brachiocephalic vein or in the upper superior caval vein, compared to placement in the lower superior caval vein or the upper right atrium.21 Catheter length was measured with arm or leg movement to make sure the tip remained in good position. We found that even a 5 millimetres change in the length makes a difference in neonates. The catheters were sutured in place as close as possible to the site of puncture of the skin to avoid movement of the catheter.
The complication rates were low. While some of that may be related to the insertion techniques previously described, the tunnel created from site of puncture to site of venous entry also likely reduced infection.22 In addition, good catheter care and maintenance by medical and nursing staff were important to help prevent the occurrence of catheter-related complications. We elected continuously to infuse heparin23 at a rate of 1 millilitre per hour (1 Units per milli-litre) into the catheter. This may decrease the risk of catheter-related formation of thrombus and, consequently, the risk of catheter-related infection.
There has been some recent concern regarding the rare, though fatal complication of myocardial or vascular perforation in neonates24. The low rate of perforation in our study may be a credit to the technique of insertion, using specific expertise, fluoroscopy and positioning the line tip no further than the cavo-atrial junction.
So far, no time limit has been established for use of peripherally inserted central venous catheters.8, 25, 26 The longest duration previously reported was 132 days, by Thiagarajan et al.8 The longest duration in our study was 139 days, when the patient had serratia marcescens positively cultured from the catheter as well as in the blood. The catheter was then removed. After appropriate intravenous antibiotic therapy, the blood culture became negative 2 days after removal, and the patient recovered successfully.
The neonates with critical congenital cardiac disease are a special population. Patients with functionally single ventricle may require bi-directional Glenn and Fontan operation in the future. In our study, peripherally inserted central venous catheters preserved the patency of the inferior and superior caval veins in such patients. We tried to avoid using the groin because of the high risk of occlusion and venous thrombosis.27 Peripherally inserted central venous catheters also preserve the femoral veins for future use. The small sample size is one of the limitations of our study. Second, the incidence rate of complications maybe was underestimated because of the retrospective nature of our study. Third, we mostly used clinical evaluations to assess thrombosis related to use of the catheters, and this may not reflect true rates. A prospective study using venography or ultrasound in all patients following removal of such catheters may be able accurately to determine the true rate of venous thrombosis associated with their use in neonates with critical congenital cardiac disease.
In conclusion, the complication rates associated with use of peripherally inserted central venous catheters were lower in neonates with critical congenital cardiac disease than in other populations. Such catheters provide reliable and safe access for prolonged intravenous therapy in neonates with critical congenital cardiac disease.
Acknowledgements
We thank all of the contributing technicians and nurses of Interventional Radiology, anaesthesiologists, as well as cardiologists and nurses of the Intensive Care Unit for their enthusiastic support and cooperation.
