Central venous access is essential to the optimal care of neonates with critical congenital heart disease (CHD). These patients have marked hemodynamic instability perinatally and perioperatively, requiring vasopressors, inotropes, blood products, antibiotics, and parenteral nutrition. While peripherally inserted central catheters (PICCs) accommodate continuous infusions, they are poorly suited to management of bleeding and hypovolemia and are associated with high rates of thrombosis (1–4). Internal jugular, subclavian, and femoral venous access can be accomplished; however, complications are more common in newborns and carry significant consequences (5–7). Thrombosis of these vessels may preclude life-saving interventions, including extracorporeal membrane oxygenation cannulation and cardiac catheterization. Most importantly in this population, patency of the superior vena cava and inferior vena cava (IVC) is essential to the successful staged palliation of single-ventricle CHD (8).
Umbilical venous catheter (UVC) insertion is the preferred vascular access approach in many cardiac centers, as it spares the aforementioned vessels (8,9). Successful UVC placement requires passing a catheter through the umbilical vein to the portal sinus (also called the umbilical recess), then through a patent ductus venosus (DV) to the subdiaphragmatic vestibule and inferior cavoatrial junction (Fig. 1). This anatomy has been well-described in the literature (10,11). DeWitt et al (12) underscored the premium placed on umbilical venous access in the perinatal management of CHD—their 2015 report details an institutional commitment to obtaining a UVC in the catheterization laboratory, if bedside attempts had failed. Even when attempted in the first hours of life, UVC insertion fails in 25–50% of newborns (13,14), most commonly because the catheter tracks into and becomes lodged in the portal veins of the liver (Fig. 2). Several techniques have been suggested to improve the success rate of the traditional blind technique, including posterior liver mobilization and dual-catheter techniques (14,15); however, the rate of failure remains higher than desired. Malpositioned UVCs should not be left in the portal circulation, due to risk of portal vein thrombosis as well as hepatic and gastrointestinal complications (11,16).
Point-of-care ultrasound (POCUS) has rapidly become standard of care for guiding arterial and central venous vascular access in infants and children, where it has been shown to improve overall and first-pass success, while reducing procedure time and complications (17–19). In a 2018 meta-analysis, de Souza et al (20) found that use of POCUS markedly reduced the rate of pediatric central venous cannulation failure, from 25% to 5.3% when compared with landmark-based techniques. Ultrasound has also been used to evaluate positioning of umbilical venous and arterial catheters tip, as reported by several groups (21–24). However, it has only very recently been described as a tool to assist during UVC placement (25). Simultaneous to this report, we have developed a related technique in our cardiac ICU (CICU) and neonatal ICU (NICU), whereby real-time procedural ultrasound guidance can be used to successfully rescue malpositioned UVCs in the majority of cases, despite failed conventional attempts by experienced providers. The aim of this project is to describe and quantify the success of this technique in three pediatric centers.
PATIENTS AND SETTING
Three CICU and NICU at tertiary children’s hospitals participated in this study. Between March 2019 and May 2021, 32 infants underwent POCUS-assisted umbilical venous cannulation across these three centers. All patients had failed a previous attempt at traditional, blind UVC insertion. Ultrasound-guided UVC procedures were performed by an attending cardiac intensivist or neonatologist credentialed in UVC placement and with extensive procedural ultrasound experience or were performed by a clinical fellow or equivalent under the direct tutelage of such an attending. Patient hospital records were reviewed, and demographics, diagnoses, and details of the umbilical vein catheterization were extracted. The study was reviewed by the Institutional Review Board at the Children’s Hospital of Philadelphia (Number 20-018237) and determined to meet exemption criteria per 45 Code of Federal Regulations 46.104(d) 4(iii) (healthcare operations).
DESCRIPTION OF TECHNIQUE
Imaging of Vasculature
Intrahepatic vasculature was imaged by ultrasound using either a General Electric Logiq P9 (GE Healthcare, Milwaukee, WI), Philips CX50 (Philips, Andover, MA) or Mindray M9 (Mindray, Mahwah, NJ), equipped with linear and phased array transducers. Most commonly, a 4–12 MHz linear array probe was employed for its wide field of view and high spatial resolution. Using the liver as an acoustic window, relevant anatomy may be visualized with the probe approximately in the midline, below the xiphoid and oriented sagittally. The umbilical vein is seen anteriorly, coursing cranially to meet the portal sinus in the liver (Fig. 3). Branches from the portal sinus include the left and right portal veins and the DV. The portal veins are identified by their echo-bright endothelium; the left veins extend anteriorly and superiorly, while the right portal veins meet the portal sinus posteriorly and inferiorly. The DV is distinguished by its oblique takeoff from the posterior aspect of the portal sinus, as it courses cranially, posteriorly, and slightly rightward to meet the hepatic veins and IVC as they enter the right atrium at the subdiaphragmatic vestibulum. To improve visualization of the portal sinus, left portal vein, and DV, one may use one or more of the following transducer movements: rotating slightly counterclockwise, sliding slightly to the infant’s left, and tilting rightward to direct the ultrasound beam toward the patient’s right side.
Confirmation of Patency of the Ductus Venosus
The patency of the DV is evaluated using color Doppler imaging at low color scale or injection of agitated blood or saline contrast (Fig. 3, B and E; Supplemental Video 1, https://links.lww.com/PCC/B996 [legend, https://links.lww.com/PCC/B1000]). Of note, a phased array probe often provides better color signal than the linear array probe preferred for B-mode visualization.
Ultrasound-Guided UVC Placement
The infant is positioned supine and the abdomen widely prepped per local unit guidance. A sterile field from hips to nipples and out to the anterior axillary line is prepared in order to allow adequate windows for imaging without compromising sterility. The patient is draped with sterile towels and the probe is placed in a sterile sleeve. If an indwelling, noncentral UVC is present, it is removed and the umbilical stump resterilized. The new UVC is prepped in standard fashion and inserted into the umbilical vein to a depth of 5–7 cm (in a term infant). At this point, it can be visualized in the umbilical vein or portal sinus (Fig. 3). Using the transducer manipulations described above, the catheter is then advanced under direct ultrasound visualization. The mechanism of malposition becomes readily apparent, as the catheter typically follows the previous, undesired course into the left portal veins anteriorly, or dives posteriorly through the portal sinus, bypassing the DV to lodge in the right portal veins (Figs. 3A and 4; Supplemental Video 2, https://links.lww.com/PCC/B997 [legend, https://links.lww.com/PCC/B1000). Unlike the traditional, blind technique, in which liver and catheter manipulations are made empirically and often unsuccessfully, ultrasound provides real-time visual feedback that allows the operator to perform informed and effective adjustments based on to the patient’s anatomy.
In order to redirect the catheter to the DV, it is retracted into the portal sinus and then readvanced while simultaneously mobilizing the liver using the operator’s imaging hand and ultrasound probe. This allows for continuous imaging during the procedure. In our experience, it is helpful to have a second provider assist with imaging and manipulation, but this is not required. During compression of the abdomen, the resulting alterations of the vascular anatomy can be visualized by ultrasound and optimized. As shown in Figure 4 and Supplemental Video 3 (https://links.lww.com/PCC/B998; legend, https://links.lww.com/PCC/B1000), posterior liver pressure distorts the portal sinus and improves the angle from the umbilical vein to the DV, allowing for a flatter, straighter trajectory. It also compresses the entrance to the left portal veins, while also making the caudal turn from the portal sinus into the right portal veins more acute and therefore less favorable. Lateral pressure away from the side of malposition may also enhance alignment. Of note, excessive compression will occlude the mouth of the DV, whereas injection of a small volume of saline through the UVC may transiently distend it and facilitate catheter engagement. Using these maneuvers, the catheter is advanced into and through the DV to the desired position at the inferior cavoatrial junction.
Difficulty may be encountered due to a small DV orifice, unfavorable takeoff angle from the portal sinus, or to the presence of additional tissue ridges and adjacent portal venous channels. In these cases, dual-catheter techniques may facilitate successful placement as demonstrated in Figure 5 and Supplemental Video 4 (https://links.lww.com/PCC/B999; legend, https://links.lww.com/PCC/B1000).
Ultrasound-guided umbilical venous cannulation was attempted in 32 cases, all following prior unsuccessful blind cannulation attempts by experienced neonatologists. Diagnoses present in these infants included transposition of the great arteries (5), pulmonary atresia or critical pulmonic stenosis (5), hypoplastic left heart syndrome (4), tetralogy of Fallot variants (4), interrupted aortic arch (3), Shone complex (3), double-outlet right ventricle (2), atrioventricular canal defect (2), total anomalous pulmonary venous connection (3), and truncus arteriosus (1). Mean (sd) gestational age and weight were 38.1 weeks (1.49 wk) and 3.00 kg (0.46 kg), respectively. Median postnatal age at time of attempted UVC replacement was 16.1 hours (interquartile range, 4.9–37.7 hr). Twenty-four patients (75%) were on prostaglandin E1 infusion for ductal-dependent cardiac physiology at the time of UVC placement. The frequency of ultrasound-assisted UVC procedures increased during the study period, with the majority being performed more recently. Specifically, 23 (72%) procedures were performed at the Children’s Hospital of Philadelphia in the last 9 months of the 26-month study period. Review of our recent institutional data demonstrated 75% success in conventional UVC placement, with a similar success rate in cardiac and noncardiac neonates. These were typically performed by neonatologists in that children’s hospital’s dedicated labor and delivery unit.
Overall, use of POCUS facilitated the rescue of malpositioned UVCs in 23 of 32 patients (72%), 20 of 25 (80%) performed in the first 48 hours of life and three of seven (43%) performed later. UVC rescue was successful in 18 of 24 patients (75%) on prostaglandin E1 at the time of the procedure, versus three of eight patients (38%) not on prostaglandin E1. Reasons for failure included small DV aperture, DV spasm following catheter instrumentation, and unfavorable angle of DV takeoff despite liver manipulation. In eight of 23 patients (35%), the UVC was somewhat deep once manipulated through the DV, based a combination of x-ray and ultrasound assessment; no catheter was too shallow. At the time of UVC placement, the mean weight was 3.05 kg (range, 2.04–3.80 kg).
In several patients where cannulation of the DV has proven difficult despite these ultrasound-augmented techniques, we have used dual-catheter techniques analogous to those used in traditional, blind placement. Using ultrasound, the mechanism of this long-established practice is readily apparent. The first catheter occupies a more facile path into a portal vein and renders it less accessible to a second catheter, which may then be directed through the DV. We have also found that in some cases, contrary to conventional teaching, a more flexible 3.5F catheter may be successful where a stiffer 5F catheter was not, as it tends to drop more easily into the posterior aspect of the portal sinus and may more easily enter a small DV. However, a 5F double-lumen UVC may be preferred for perioperative management based on the expected needs for fluid resuscitation and medication infusions.
UVCs had a mean dwell time of 165.5 hours (73.9 hr) and were used for vasoactive medications, blood products, and parenteral nutrition. No infant required vascular access devices in the internal jugular, subclavian, or femoral veins while the UVC was in situ. There were no procedural or catheter-associated complications, including liver or vascular injury, bacteremia, or omphalitis. There were no reported cases of portal vein thrombosis.
In this study, we demonstrate that real-time POCUS may be used to achieve optimal UVC positioned with a high rate of success, even after unsuccessful attempts by experienced providers using traditional techniques. Using ultrasound, the provider can make adjustments that are impossible without continuous visual feedback, correcting the catheter course and manipulating the intrahepatic vascular anatomy to achieve the desired trajectory.
Reported rates of successful UVC placement using traditional “blind” techniques range from 50% to 75% (13,14), which is consistent with a recent internal review of our Special Delivery Unit experience at the Children’s Hospital of Philadelphia. We have demonstrated that it is possible to salvage malpositioned UVCs and achieve central venous access in most patients using real-time ultrasound guidance. This technique has the potential to greatly improve the overall success of umbilical venous cannulation, from 50% to 75% to approximately 90–95%. This is similar to the improvement imparted by POCUS to other types of central venous access in children (20). Our experience supports an emphasis on attempting UVC placement early after birth to maximize the chances of DV patency. The additional manipulation does not appear to be associated with increased risk of procedural or catheter-related complications.
Successful implementation of this technique depends upon the availability of appropriate equipment and personnel. While POCUS in the pediatric emergency department has a longer track record, utilization of ultrasound by bedside providers in the CICU, NICU, and PICU is a more recent trend. For most infants with complex CHD, definitive therapy takes place in large tertiary children’s hospitals, which commonly have the proper ultrasound equipment needed for this technique. Similarly, many NICUs have POCUS capabilities and emerging handheld devices may make bedside imaging accessible in an even larger number of centers. As the technique may require additional time to complete, we do not recommend that it be used in emergent situations or during resuscitation. We also discourage attempted catheter replacement if pre-procedure ultrasound imaging suggests that the DV is no longer patent. However, if employed early and more routinely in the course of nonemergent UVC placement, the described technique may reduce both time and number of x-ray images required, as has been shown to be the case in PICC line insertion (26). Additionally, successful placement of a central UVC should reduce both the risks of malpositioned catheters as well as the costs associated with additional vascular access procedures in neonates needing only short-term central venous access, although this remains to be proven.
The patients in this study all had CHD, and most were term or late preterm. The results may therefore not be generalizable to premature infants. In addition, the majority of patients were receiving prostaglandin therapy at the time of UVC placement. The DV closes postnatally beginning at the umbilical end, typically within the first 4 days of life in term infants and within the first 6 days in premature neonates (27). There is some evidence that prostaglandin E1, prostaglandin E2, and prostaglandin I2 act to relax the DV, whereas prostaglandin endoperoxide analogs, prostaglandin F2a, and corticosteroid exposure cause constriction (27,28). Therefore, it is possible that prostaglandin E1 therapy used for ductal-dependent CHD may extend DV patency and increase the success of this procedure.
The small number of patients in this study precludes statistical analysis of factors impacting success versus failure and also limits our ability to assess the rate of complications. It is also possible that delayed replacement may result in more complications than have been appreciated in this small study due to a decreased ability to sterilize the umbilical stump. Additionally, the three institutions collaborating on this work are academic pediatric care centers in which providers had significant prior experience using procedural and diagnostic point-of-care ultrasound. In this setting, trainees with procedural ultrasound and vascular access experience quickly oriented to the technique and were able to perform it successfully under the supervision of one of the attending coauthors. Providers without such experience will likely require additional time to learn the described techniques, potentially limiting the generalizability to other practitioners.
In considering a new procedure, it is important to ensure that adequate patient safeguards are in place and that institutional policies regarding new procedures and interventions are followed. In our centers, critical care proceduralists undergo formal hospital credentialing, didactic and hands-on training in procedural and diagnostic ultrasound, and participate in ongoing quality assurance efforts. Fellows undergo ultrasound training as part of their educational curriculum and are supervised by ultrasound-proficient attending physicians. With respect to ultrasound-guided umbilical venous cannulation, we have established a core group of cardiac intensivists and neonatologists to support increased utilization and are in the process of training additional advanced practice providers to act as resources in the CICU and NICU. Additionally, we are adding UVC datasets to our neonatal ultrasound simulators (EchoCom GmbH, Leipzig, Germany) to improve familiarity and proficiency with the relevant transhepatic windows. We recommend that other centers seeking to expand their POCUS capabilities follow a similarly formal and structured process.
We did not track the time required for ultrasound-guided UVC replacement attempts, so are unable to report that here. Ultrasound guidance may, if used routinely, reduce time and x-ray exposure. Although serial radiographs were not required during placement, a confirmatory x-ray was performed at the conclusion of each procedure as this is the standard for documenting line placement in our institutions. In most cases, there was reasonably good correlation between x-ray and ultrasound assessment of line position. Ultrasound was used as the primary modality for determining tip position, its accuracy having been well-described in the literature (21–24). However, several catheters were determined to be deeper than desired and retracted. This may reflect operators’ prioritization of secure, central placement, preferring to err on the side of intracardiac position rather than risk a shallow, noncentral position, or dislodgement.
This multicenter case series demonstrates a novel POCUS technique for obtaining umbilical venous access in neonates. Using real-time ultrasound guidance, the success rate of umbilical venous cannulation can be markedly increased, satisfying the access requirements of neonates with critical CHD without compromising the central veins that may be essential for later surgical and transcatheter intervention.
We thank Tess Marhofer for her illustrations of the umbilical venous anatomy and catheter course.
1. Barone G, D’Andrea V, Vento G, et al.: A systematic ultrasound evaluation of the diameter of deep veins in the newborn: Results and implications for clinical practice. Neonatology. 2019; 115:335–340
2. Gnannt R, Waespe N, Temple M, et al.: Increased risk of symptomatic upper-extremity venous thrombosis with multiple peripherally inserted central catheter insertions in pediatric patients. Pediatr Radiol. 2018; 48:1013–1020
3. Nifong TP, McDevitt TJ: The effect of catheter to vein ratio on blood flow rates in a simulated model of peripherally inserted central venous catheters. Chest. 2011; 140:48–53
4. Thornburg CD, Smith PB, Smithwick ML, et al.: Association between thrombosis and bloodstream infection in neonates with peripherally inserted catheters. Thromb Res. 2008; 122:782–785
5. DiPietro LM, Gaies M, Banerjee M, et al.: Central venous catheter utilization and complications in the pediatric cardiac ICU: A report from the Pediatric Cardiac Critical Care Consortium (PC4). Pediatr Crit Care Med. 2020; 21:729–737
6. Aiyagari R, Song JY, Donohue JE, et al.: Central venous catheter-associated complications in infants with single ventricle: Comparison of umbilical and femoral venous access routes. Pediatr Crit Care Med. 2012; 13:549–553
7. Costello JM, Clapper TC, Wypij D: Minimizing complications associated with percutaneous central venous catheter placement in children: Recent advances. Pediatr Crit Care Med. 2013; 14:273–283
8. Perry T, Ullman AJ, Aiyagari R, et al.: The Michigan appropriateness guide for intravenous catheters in children with congenital heart disease: miniMAGIC-CHD. Cardiol Young. 2021; 31:1814–1818
9. Ullman AJ, Bernstein SJ, Brown E, et al.: The Michigan appropriateness guide for intravenous catheters in pediatrics: miniMAGIC. Pediatrics. 2020; 145:S269–S284
10. Mavrides E, Moscoso G, Carvalho JS, et al.: The anatomy of the umbilical, portal and hepatic venous systems in the human fetus at 14-19 weeks of gestation. Ultrasound Obstet Gynecol. 2001; 18:598–604
11. Oestreich AE: Umbilical vein catheterization–appropriate and inappropriate placement. Pediatr Radiol. 2010; 40:1941–1949
12. DeWitt AG, Zampi JD, Donohue JE, et al.: Fluoroscopy-guided umbilical venous catheter placement in infants with congenital heart disease. Congenit Heart Dis. 2015; 10:317–325
13. Haase R, Hein M, Thale V, et al.: Umbilical venous catheters - analysis of malpositioning over a 10-year period. Z Geburtshilfe Neonatol. 2011; 215:18–22
14. Pennaforte T, Klosowski S, Alexandre C, et al.: Increased success rate in umbilical venous catheter positioning by posterior liver mobilization. Arch Pediatr. 2010; 17:1440–1444
15. Mandel D, Mimouni FB, Littner Y, et al.: Double catheter technique for misdirected umbilical vein catheter. J Pediatr. 2001; 139:591–592
16. Morag I, Epelman M, Daneman A, et al.: Portal vein thrombosis in the neonate: Risk factors, course, and outcome. J Pediatr. 2006; 148:735–739
17. Oulego-Erroz I, González-Cortes R, García-Soler P, et al.; RECANVA collaborative study: Ultrasound-guided or landmark techniques for central venous catheter placement in critically ill children. Intensive Care Med. 2018; 44:61–72
18. Kantor DB, Su E, Milliren CE, et al.: Ultrasound guidance and other determinants of successful peripheral artery catheterization in critically ill children. Pediatr Crit Care Med. 2016; 17:1124–1130
19. Fraga MV, Stoller JZ, Glau CL, et al.: Seeing is believing: Ultrasound in pediatric procedural performance. Pediatrics. 2019; 144:e20191401
20. de Souza TH, Brandao MB, Nadal JAH, et al.: Ultrasound guidance for pediatric central venous catheterization: A meta-analysis. Pediatrics. 2018; 142:e20181719
21. Ades A, Sable C, Cummings S, et al.: Echocardiographic evaluation of umbilical venous catheter placement. J Perinatol. 2003; 23:24–28
22. Fleming SE, Kim JH: Ultrasound-guided umbilical catheter insertion in neonates. J Perinatol. 2011; 31:344–349
23. Michel F, Brevaut-Malaty V, Pasquali R, et al.: Comparison of ultrasound and x-ray in determining the position of umbilical venous catheters. Resuscitation. 2012; 83:705–709
24. Simanovsky N, Ofek-Shlomai N, Rozovsky K, et al.: Umbilical venous catheter position: Evaluation by ultrasound. Eur Radiol. 2011; 21:1882–1886
25. Kishigami M, Shimokaze T, Enomoto M, et al.: Ultrasound-guided umbilical venous catheter insertion with alignment of the umbilical vein and ductus venosus. J Ultrasound Med. 2020; 39:379–383
26. Katheria AC, Fleming SE, Kim JH: A randomized controlled trial of ultrasound-guided peripherally inserted central catheters compared with standard radiograph in neonates. J Perinatol. 2013; 33:791–794
27. Kondo M, Itoh S, Kunikata T, et al.: Time of closure of ductus venosus in term and preterm neonates. Arch Dis Child Fetal Neonatal Ed. 2001; 85:F57–F59
28. Adeagbo AS, Coceani F, Olley PM: The response of the lamb ductus venosus to prostaglandins and inhibitors of prostaglandin and thromboxane synthesis. Circulation Res. 1982; 51:580–586