Secondary Logo

Journal Logo

Intensivists Performed Percutaneous Bicaval Double-Lumen Echo-Guided Extracorporeal Membrane Oxygenation Cannulation at Bedside in Newborns and Children

A Retrospective Analysis

Moscatelli, Andrea, MD1; Febbo, Francesca, MD2; Buratti, Silvia, MD1; Pezzato, Stefano, MD1; Bagnasco, Francesca, BS3; Lampugnani, Elisabetta, MD1; Nuri, Halkawt, MD4; Buffelli, Francesca, MD5; Grasso, Chiara, MD6; Castagnola, Elio, MD7

Pediatric Critical Care Medicine: June 2019 - Volume 20 - Issue 6 - p 551–559
doi: 10.1097/PCC.0000000000001918
Extracorporeal Support

Objectives: To describe a single institution experience on echo-guided percutaneous bicaval double lumen extracorporeal membrane oxygenation cannulation performed at the bedside by intensivists.

Design: Retrospective observational study.

Setting: Extracorporeal membrane oxygenation team of a tertiary care children’s hospital.

Patients: All patients 0–14 years old undergoing venovenous extracorporeal membrane oxygenation from January 1, 2013, to January 1, 2018.

Interventions: None.

Measurements and Main Results: Thirty children underwent 32 extracorporeal membrane oxygenation runs. Median age at enrollment was 2 months (interquartile range, 0–20.5 mo), 65.6% of the runs (21 patients) were performed in newborns (n = 13, 40.6%) or infants (n = 8, 25%). Median preextracorporeal membrane oxygenation index was 66.9 (interquartile range, 50–85.6). Major comorbidities were present in 50% of patients. All patients were cannulated percutaneously. In two cases cannulation occurred from the left internal jugular vein. Extracorporeal membrane oxygenation was effective in increasing pH, arterial oxygen saturation, PaO2, and lowering PaCO2. The overall differences in pre and postextracorporeal membrane oxygenation values were statistically significant, while stratifying patients according to the cannula diameter (mm)/major diameter of the cannulated internal jugular vein (mm) ratio (> 0.67 or ≤ 0.67), statistical significance was reached only for the highest ratio. Complications were observed in three runs: two cannula tip dislocations in the right atrium and one limited flow in the only case in which an Avalon cannula was not used. In 20 cases (62.5% of 32 runs), the cannulated vessel was patent at follow-up or autopsy. A ratio less than or equal to 0.67 or greater than 0.67 did not influence the occurrence rate of complications, nonpatency of the internal jugular vein, death for intracranial bleeding and death at 30 days from extracorporeal membrane oxygenation discontinuation. Overall cumulative survival at 30 days from extracorporeal membrane oxygenation discontinuation was 60% (95% CI, 40–75), with a survival advantage in the case of ratio greater than 0.67 (65%; 95% CI, 44–80 vs 25%; 95% CI, 0–60).

Conclusions: The described technique proved to be feasible, safe, and effective. Further investigation is needed.

1Neonatal and Pediatric Intensive Care Unit, Department of Surgery and Critical Care, IRCCS Istituto Giannina Gaslini, Genova, Italy.

2Postgraduate School in Anesthesia and Intensive Care, University of Genova, Genova, Italy.

3Epidemiology and Biostatistics Unit, IRCCS Istituto Giannina Gaslini, Genova, Italy.

4Division of Cardiovascular Surgery, Department of Surgery and Critical Care, IRCCS Istituto Giannina Gaslini, Genova, Italy.

5Fetal-Perinatal Pathology Unit, Mother and Child Department, IRCCS Istituto Giannina Gaslini, PhD Course in Pediatric Sciences, Fetal-Perinatal and Pediatric Pathology, University of Genova, Genova, Italy.

6Postgraduate School in Anesthesia and Intensive Care, University of Catania, Catania, Italy.

7Division of Infectious Diseases, Department of Pediatrics, IRCCS Istituto Giannina Gaslini, Genova, Italy.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (

The authors have disclosed that they do not have any potential conflicts of interest.

Address requests for reprints to: Andrea Moscatelli MD, Neonatal and Pediatric ICU, Department of Surgery and Critical Care, IRCCS Istituto Giannina Gaslini, Genova, Italy. E-mail:

Venovenous extracorporeal membrane oxygenation (ECMO) should be the preferred configuration for respiratory extracorporeal life support (ECLS), given the relevant complications associated with venoarterial cannulation: neurologic and distal limb morbidity, embolism, perfusion of the lungs and coronary arteries with poorly oxygenated blood, and increased afterload of the left ventricle (1, 2). ECMO related neurologic complications are reported in 10–20% of cases and remain a major cause of death and long-term morbidity in newborns and children. Venoarterial ECMO is one of the major risk factors (3–5). Venovenous double lumen cannulation showed increasing popularity over years, accounting for 20.2% of neonatal and pediatric cases (6). Single vessel cannulation by the right internal jugular vein (RIJV) or left internal jugular vein (LIJV) overcomes the limitations related to the small dimensions of the femoral veins in newborns and infants. Other advantages include faster implementation of ECLS in case of emergencies, lower risk of infections, and easier mobilization and participation in physical therapy (7, 8). Furthermore, newer cannulas like the Avalon Elite (Maquet, Rastatt, Germany) have improved effectiveness with their exclusive design. The aspiration ports are located in the superior vena cava (SVC) and inferior vena cava (IVC), whereas the infusion port faces the tricuspid valve (TV) (Fig. 1). In conjunction with wire reinforcement which prevents catheter bending, this allows a constant and unobstructed flow even with centrifugal pumps while minimizing recirculation (9, 10).

Figure 1

Figure 1

Data on true percutaneous echo-guided bicaval double lumen (BCDL) ECMO cannulation performed at the bedside by intensivists are scant in pediatrics, especially in newborns and infants (5). In this patient population, the venoarterial configuration is still the most common for respiratory support (70.9%) (11). Results on BCDL cannulation in pediatrics are contradictory, and outcomes were encouraging in some series (9, 10, 12), whereas major complications and cannula migration problems were reported in others (13, 14). The best insertion technique has not yet been defined. The aim of this study is to describe our experience on BCDL cannulation at the bedside in newborns and children, evaluating its feasibility, effectiveness, and safety.

Back to Top | Article Outline


Study Design

We conducted a retrospective observational study on all patients 0–14 years old who underwent venovenous ECMO cannulation by our ECMO team, from January 1, 2013, to January 31, 2018. Follow-up was closed on August 30, 2018. Informed consent was obtained from parents before ECMO initiation and data was treated in compliance with the actual European Union regulations. This study is waived from Institutional Review Board approval. For each ECMO run, the following data were collected: patients’ demographics (age, gender, weight), diagnosis, comorbidities, main indication to ECMO, pre-ECMO ventilation mode, ECMO duration, pre-ECMO oxygenation index (OI = mean airway pressure × PaO2/FIO2), pre- and post-ECMO blood gas analysis values (pH, PaCO2, PaO2, arterial oxygen saturation [SO2]), ECMO flow rates, revolutions per minute (RPM), venous aspiration pressures (mm Hg), cannula size (F), type and insertion site, IJV dimensions (mm), ratio between the cannula diameter (DC) in mm and the major diameter of the cannulated IJV (DIJV, mm), ratio between the DC in F and patients’ weight (kg), complications, IJV patency at ultrasound follow-up (at 30 d from ECMO discontinuation) or at autopsy, death at 30 days from ECMO weaning, death to discharge, and cause of death. Complications were defined as heart or great vessels perforation, pericardial effusion/cardiac tamponade, death directly correlated to the cannulation procedure, bleeding from the cannula insertion site requiring transfusion of 10 mL/kg or greater of packed RBCs, infection of the cannula insertion site, cannula dislocation (tip in right atrium [RA], obstructive in an hepatic vein, or ductus venosus, too deep in the IVC), cannula thrombosis, obstructed flow, recirculation, and need for cannula site repair. In adult patients, it is common practice not to use cannulas greater than 2 of 3 the vessels diameter (i.e., DC/DIJV ≤ 0.67) (15), whereas in pediatrics, to prevent cannulation related vascular injuries, a careful ultrasound assessment of the IJV dimensions has been suggested (10). Furthermore, a DC/Weight greater than 0.4 has been associated with a higher risks of vascular thrombosis in children after central venous line placement (16). Thus, data has also been analyzed according to these ratios.

Back to Top | Article Outline

Cannulation Technique and Patients’ Management

Patients were placed on ECMO according to the judgment of the attending physician on call if presenting refractory hypoxia and/or hypercapnia not responding to maximized medical therapy and with no contraindications to ECMO. Cannulations were performed at bedside in the ICU by a team of two intensivists and one medical staff in training for cannulation or a scrub nurse. Transthoracic echo (TTE) control was provided by one of the two senior intensivists under sterile conditions (SonoSite M-Turbo, Bothell, WA; pediatric cardiac probe 4–8 MHz). The surgical team was alerted and readily available in case of major complications. After sedation, paralysis, antibiotic prophylaxis, and skin disinfection, the IJV was measured with echo in all patients (major and minor diameter, mm) at the level of the expected cannula insertion site, which was located at mid-length between the mandibular angle and the clavicle (SonoSite M-Turbo; hockey stick probe, 6–13 MHz). Ideally, this could allow the cannula to have a sufficient subcutaneous course before entering the chest. A projection of the cannula length on a standard chest radiograph (from estimated insertion site to proximal IVC) was drawn in order to approximately predict that the aspiration ports would be properly located once the cannula was in place (10). A more distal insertion site was chosen if the cannula was suspected to be too short to reach the IVC with its tip. The IJV was accessed with a 20 gauge cannula under off-plane echo direct visualization. A 0.64 mm guidewire was passed through the 20 gauge cannula, allowing the introduction of a 4F (5 cm length) hemostasis valve introducer (Terumo, Shibuya, Tokyo, Japan). The introducer can accept a 0.97 mm guidewire suitable for the tapered dilators of the Avalon Elite introduction kits. This cannula is our first choice for venovenous ECMO support and is available in the 13F, 16F, 19F, 20F, 23F, 27F, and 31F sizes. The cannula size was chosen according to the predicted needed flows. The IVC was insonated on a subcostal long axis view, to identify its course and the origin of the hepatic veins and ductus venosus in newborns (Fig. 2, A and B). The straight and uncoiled progression of the 0.97 mm guidewire, deeply into the IVC, had to be mandatorily followed and displayed under real-time TTE control (subcostal RA and four-chambers views, subcostal bicaval and long axis IVC views) (Fig. 2C–I). Since the SVC and the IVC lie on different planes (Figs. 1 and 2), positioning of the guidewire in the IVC was facilitated using a straight guide with the extremity bent in a hockey stick shape at the same angulation as the one between the two vessels (0.97 mm diameter fixed core wire guide, straight; Cook Medical, Bloomington, IN; 145 cm length for the 13–19F cannulas and 260 cm for the 20–31F) (Fig. 2, C and E). Presence of guidewire loops and bending in the RA and right ventricle should be excluded by cardiac echo (subcostal four-chambers view) and suspected in case of extrasystoles on the electrocardiogram, rhythm changes, or cardiovascular instability. Looping can lead to cardiac perforation during dilation and cannula positioning (17). Serial gentle dilations were performed under continuous echo visualization, always displaying the straightness of the guidewire and its correct placement in the IVC. Excessive incision or dilation of the skin was avoided to prevent insertion site bleeding. Once the guidewire was in place and dilation completed and no contraindications to anticoagulation were present, a dose of heparin up to 50–100 U/kg was given and titrated according to the basal activated coagulation time, with a target of 200–250 seconds. After flushing with sterile saline, the cannula was inserted in place under real-time TTE control. Using the infusion port to tip distance as a reference, which varies according to catheter dimension (13F: 2.8 cm, 16F: 4 cm, 19F: 5.7 cm, 20F–23F–27F–31F: 9.4 cm), and verifying that the tip was not displaced in and obstructing a hepatic vein or ductus venosus (Fig. 1C), optimal positioning with the infusion port facing the TV was obtained with echo. The patient was then connected to the ECMO circuit (Levitronics Pedivas/Centrimag pump, Thoratec Corporation, Pleasanton, CA). After placing the head in a neutral position and re-checking the cannula with echo, fixation was provided with at least three 2-0 silk ties and completed with an occlusive dressing. To prevent pressure sores, DuoDERM dressing (ConvaTec, Deeside, Flintshire, United Kingdom) was placed between the skin and the cannula. Ventilation was set with resting parameters aimed at optimizing lung recruitment with a 0.35–0.40 FIO2 (18). To exclude any displacement of the cannula, echo was repeated every 8 hours and in case of unexpected reductions in flows or increase in the negative aspiration pressures (12). ECLS was considered to be effective when the following therapeutic goals were met: 7.35 < pH < 7.45, 40 mm Hg < PaCO2 < 45 mm Hg, 87% < arterial SO2 < 97%, 50 mm Hg < PaO2 < 90 mm Hg, with an improvement respect to the pre-ECMO values. Minimum recommended flow rates were 120 mL/kg/min (newborns), 100 mL/kg/min (infants), 80–90 mL/kg/min (toddlers-children) with aspiration pressures not lower than –100 mm Hg (19, 20).

Figure 2

Figure 2

In patients with unfavorable outcomes undergoing autoptic examination, the protocol included a careful dissection of the IJV, innominate vein, SVC, RA, and IVC (Fig. 1) in order to exclude infection of the cannula insertion site, thrombosis, and vascular or heart cannulation related injuries.

Back to Top | Article Outline

Statistical Analysis

Descriptive statistics were reported in terms of absolute frequencies and percentages for qualitative data. The Pearson’s chi-square test or Fisher exact test, if appropriate, were applied to compare proportions. Quantitative data were described in terms of median values, range, and interquartile range (IQR) due to their nonnormal distribution. Accordingly, comparisons between groups were made by the nonparametric Mann-Whitney U test or Kruskal-Wallis test. The intrapatient blood gas analysis values (pre- and post-ECMO) were compared using the Wilcoxon signed rank test. The cumulative survival probability was calculated by the Kaplan-Meier method at 30 days from ECMO discontinuation, and the 95% CI of the estimates were calculated according to the Kalbfleisch and Prentice formula. All tests were two-tailed and a p value of less than 0.05 was considered statistically significant. All analyses were performed using Stata Statistical Software, Release 13.1 (StataCorp LLC, College Station, TX).

Back to Top | Article Outline


Patients’ Characteristics

Thirty patients, 96.7% 0–7 years old, were cannulated in 32 different occasions, two of them undergoing two ECMO runs (both 3.4 kg infants). Characteristics of the ECMO runs are reported in Table 1. Twenty-one (65.6%) runs were performed in newborns and infants (five infants weighing < 5 kg). Median gestational age and weight of the newborns were 38 weeks (range, 36–40 wk) and 3.3 kg (range, 1.6–4.2 kg), respectively. Acute hypoxic respiratory failure (AHRF) was the most frequent indication to ECMO (29 runs, 90.6%), with a median pre-ECMO OI of 66.9 (IQR, 50–85.6). Only two patients required ECLS for severe hypercapnia and one as a bridge to lung transplantation. Acute respiratory distress syndrome was the prevailing diagnosis, followed by perinatal causes of AHRF: congenital diaphragmatic hernia (CDH), persistent pulmonary hypertension of the newborn, and meconium aspiration syndrome. Significant comorbidities were present in 50% of the runs (Table 1). Thirteen patients were assisted on high-frequency oscillatory ventilation with inhaled nitric oxide prior to ECMO initiation. Almost all the patients were cannulated with a BCDL Avalon catheter, size ranging from 13F to 27F. In only one case, due to inadvertent contamination of the cannula during positioning, a spare double-lumen OriGen Reinforced 13F cannula (OriGen Biomedical, Austin, TX) was inserted because of unavailability of a second Avalon. In all age groups, every cannulation was performed using the described percutaneous technique under TTE control. The insertion site in 30 cases was the RIJV. Two cases, one newborn (3.8 kg) and one infant (3.4 kg) respectively, were cannulated through LIJV since the RIJV was not patent due to previous central venous access and ECMO cannulation. The backup surgical team was never activated. Three infants were cannulated at a different institution by our ECMO retrieval team. One patient was successfully percutaneously converted to veno-veno-arterial (VVA) ECMO by adding an arterial cannula with a Y connection to the reinfusion limb of the circuit. Median ECMO duration was 10 days (IQR, 6.5–18.5 d) with a maximum of 49 days.



Back to Top | Article Outline

Effectiveness of ECMO Support

ECLS was effective in increasing pH, arterial SO2, PaO2, and lowering PaCO2. The overall differences in pre- and post-ECMO values were statistically significant, whereas stratifying patients by DC/DIJV ratio (> 0.67 or ≤ 0.67) statistical significance was reached only for the highest ratio. Nevertheless, pH, PaCO2, arterial SO2, and PaO2 ECLS goals were always met in both groups (Supplemental Table 1, Supplemental Digital Content 1,

Post-ECMO PaO2, post-ECMO arterial SO2, and ECMO flow rates did not differ according to age groups. RPM resulted higher and venous aspiration pressures were more negative in toddlers. Minimum age recommended flow rates and oxygenation goals were met in all age groups. Flow rates were well above the recommended targets, especially in older patients. Negative aspiration pressures remained within safety limits in all the subgroups (Table 2). Stratifying data according to cannula size: age, weight, and DC/weight ratios obviously varied with statistical significance, whereas no difference resulted in terms of flow rates, RPM, venous aspiration pressures, DC/DIJV, complications, and nonpatency of the IJV (Table 3).





Back to Top | Article Outline

Complications and Mortality

Overall complications were observed in three ECMO runs (9.4%), all newborns: two cannula tip dislocations in the RA, and one obstructed/limited flow with the OriGen cannula (median flow 80 mL/kg/min; IQR, 72.5–82.5). Patency data was available on a total of 24 runs. In 20 of those runs (83.3%), the cannulated vessel was patent at follow-up (n = 15) or autopsy (n = 5). Thirteen patients (43.3%) died due to irreversible lung disease (n = 7) (worsening and progressive parenchymal disease after at least 2 wk of ECMO support, judged with no potential for recovery and with no indication to lung transplantation) (21), brain hemorrhage (n = 5), and one for congenital heart disease not amenable of surgical correction (56.7% survival to discharge). Overall, cumulative survival at 30 days from ECMO discontinuation was 60% (95% CI, 40–75).

According to DC/DIJV ratio (≤ 0.67 or > 0.67), there was not a different distribution of: frequency of complications (1 for rate ≤ 0.67 vs 2 for rate > 0.67; p = 0.410); nonpatency of the IJV (0 for rate ≤ 0.67 vs 4 for rate > 0.67; p = 1.000); overall death at 30 days from ECMO discontinuation (3 for rate ≤ 0.67 vs 9 for rate > 0.67; p = 0.338); and death by intracranial bleeding (1 for rate ≤0.67 vs 4 for rate >0.67, p = 1.000). Cumulative survival estimates showed an increased survival rate at 30 days from ECMO discontinuation in patients with a DC/DIJV greater than 0.67 (65%; 95% CI, 44–80 vs 25%; 95% CI, 0–60) (Fig. 3). Autoptic examinations carried out in five patients did not find any sign of cannulation related injuries, thrombosis (vessels and heart), or infection (cannula insertion site). In one patient, a thin plaque of fibrinous deposition on the endothelial surface of the SVC and IVC at the level of the aspiration ports was detected (Supplemental Fig. 1, Supplemental Digital Content 2,; legend: Fibrinous depositions on the endothelial surface of the superior and IVC [black arrows] in correspondence with the aspiration ports. Possible effect of mechanical stimulation on the intimal layer, depositions were tenacious with no risk of embolization.).

Figure 3

Figure 3

Back to Top | Article Outline


This is probably one of the largest pediatric studies on BCDL cannulation performed percutaneously at the bedside by intensivists under exclusive TTE guide. The most recent analysis of the Extracorporeal Life Support Organization (ELSO) registry reports only 188 percutaneous cannulations below 1-year-old versus 1,117 surgical cutdowns, with no mention of neonatal cases (5). Most of our runs were performed in newborns, infants, or in patients weighing less than 5 kg, conditions, which were reported to be a contraindication to percutaneous cannulation in other series (10). Data on the relationship between cannula size and IJV dimensions and on vessels’ patency after decannulation are reported, which we did not find previously described in pediatrics. The technique proved to be safe, effective, and suitable for ECMO retrievals from peripheral institutions, where access to the operatory room and fluoroscopy can be difficult. In our 32 consecutive cases, we did not find any contraindications to the percutaneous approach. We would have considered a surgical cutdown if the IJV was not accessible with the 22 gauge cannula or in case of hematoma or vascular laceration. IJV dimensions, even if recorded, never influenced the size of the chosen cannula. Real-time ultrasound guidance plays a crucial role to prevent vascular or cardiac injuries and to advance the guidewire into the IVC. The J tip guidewires included in the Avalon introduction kits are less effective for this purpose, and custom bent ones have been used instead (see Cannulation technique and patient’s management). No attempts at dilation or cannula progression have to be made unless the guidewire is proven to be straight and uncoiled from the SVC to the IVC (Fig. 2). Furthermore, with echo, it is possible to precisely position the infusion port right in front of the TV, limiting recirculation to the minimum (Fig. 1). With the exclusion of the case in which we used the OriGen cannula, we never experienced problems of limited flow, probably due to the peculiar bicaval design of the Avalon Elite and to the fine-tuning in ultrasound-guided positioning (Tables 2 and 3). The rate of cannula malpositions/flow problems and repositioning was extremely low compared with other studies (14): respectively 9.4% versus 52% and 6.3% versus 40%, even if echo was the only imaging used to guide cannulation. Flows were always adequate to ensure oxygenation and CO2 removal with safe aspiration pressures (Supplemental Table 1, Supplemental Digital Content 1,; and Tables 2 and 3). Most likely, the frequent ultrasound monitoring (at least every 8 hr) to optimize patient position and to detect early signs of cannula displacement was responsible for these low rates. Accidental dislocation of the tip of the cannula into the RA occurred in only two newborns during mobilization within the first 48 hours from ECMO initiation. Repositioning without circuit discontinuation was easy under TTE control by aligning the main axis of the SVC with the axis of the IVC by head flexion, back elevation, and rotating the cannula during advancement in the IVC if needed. We speculate that, with time, the cannula adapts to the vascular structures making migration less probable. Unlike Speggiorin et al (13), we never retracted the tip of the Avalon Elite into the RA to improve flow nor would have ever allowed such a maneuver. This was likely related to the high rate of atrial perforations reported in their study (6.9%). In several cases, we allowed the tip of the cannula to sit at the confluence between the ductus venosus or a hepatic vein with the IVC instead, provided that venous flow from the liver was not hampered and that the cannula was not displaced too deep into the IVC (12).

The rationale behind not using an oversized cannula proportionate to the vessel’s dimensions (DC/DIJV ≤ 0.67, Dc/weight ≤ 0.4) is to prevent vascular lacerations during cannulation, drainage problems, and thrombosis (10, 15, 16). In our series, median DC/DIJV was 0.8 (0.7–0.9 IQR), and no patients had a Dc/weight less than or equal to 0.4 (median, 3.2; IQR, 1.7–3.8). Nevertheless, we did not experience any vascular or heart injuries, no signs of thrombosis, or vascular iatrogenic lesions were detected at postmortem examinations, and the vascular axis was patent at follow-up or autopsy in a total of 20 cases. No deaths were secondary to cannulation. Interestingly, improvement in blood gas analysis values after ECMO was statistically significant only in the group of patients with DC/DIJV greater than 0.67 (Supplemental Table 1, Supplemental Digital Content 1,, suggesting that a high ratio does not affect cannula functionality. This observation is corroborated by the fact that frequency of complications, nonpatency of the IJV, mortality at 30 days, and mortality for intracranial bleeding were not different according to DC/DIJV less than or equal to 0.67 or greater than 0.67. We observed an increased survival rate at 30 days in patients with DC/DIJV greater than 0.67 (Fig. 3). Furthermore, cannula size did not influence the rate of complications or IJV occlusion (Table 3). In newborns and children, cannulas capable of ensuring adequate flows are relatively large compared with IJV dimensions, and it is virtually impossible to respect the above-mentioned criteria (i.e., DC/DIJV ≤ 0.67 and Dc/weight ≤ 0.4) while warranting an effective ECLS. Respecting them, we probably would not have cannulated any patient. Therefore, we think that the choice of the cannula size should be made taking into account a careful balance between patients’ flow needs and IJV dimensions. The relative importance of either of the two factors has to be established case by case. A percutaneous technique allows safe insertion of large catheters in small vessels since the surrounding tissues provide support to the veins’ walls, which can be distended without tearing during serial dilations. In larger patients, “stiff” guidewires can be useful to prevent looping and trans-esophageal echo should be taken into consideration if the acoustic window is not optimal (22). Two patients weighing less than 4 kg were successfully cannulated from the LIJV. We did not find any previous reports on this ECMO vascular access in newborns and infants.The rate of intracranial bleeding was 15.6% (on 32 runs), a higher percentage compared with the ELSO registry for neonatal (12.3%) and pediatric (6.9%) respiratory ECMO. Survival to discharge was 69.2% in newborns and 47.1% in pediatric patients (ELSO registry benchmark for venovenous ECMO, respectively, 76% and 67%). Of note, one of our pediatric patients with an unfavorable outcome was converted to VVA-ECMO (ELSO registry survival rate 36%) (11). Analyzing our outcomes it has to be taken into consideration that we selected a high-risk population. Sixteen patients (50% of the 32 runs) presented major comorbidities; six had hematology-oncology conditions, among whom four were hematopoietic stem cell transplant (HSCT) recipients. Five patients had CDH, and one patient had severe neonatal pulmonary hypoplasia. Those developing intracranial hemorrhage were mainly hematology-oncology patients (n = 3) with coagulation and platelets disorders, one with posterior reversible encephalopathy syndrome and one CDH newborn with signs of preexisting prenatal intracranial bleeding (Table 1). All patients with brain hemorrhage died. Among the patients who died, three were HSCT recipients, two had a hematology-oncology condition, and two had CDH. According to data from the ELSO registry, only 36% of the patients with malignancies placed on ECMO for respiratory support survived to discharge, as did 5.9% of those after HSCT placed on ECMO for the same reason. Survival to discharge in newborns with CDH is 51% (11, 23, 24). The study has several limitations since it is retrospective and observational on a limited number of patients and data did not have a homogeneous distribution among groups. Definitive conclusions cannot be drawn given the relatively small sample size, a common weakness of pediatric ECMO studies; similar numbers of patients are described in the available published reports on this topic.

Back to Top | Article Outline


Our preliminary data shows that BCDL ECMO cannulation performed at the bedside by intensivists under exclusive TTE control can be feasible, effective, and safe in pediatrics. Apparently, IJV dimensions should not be the only factor influencing the choice of the cannula size, and priority should be given to the potentially more effective one to warrant adequate flows. The LIJV approach might be taken into consideration if the RIJV is not practicable.

We believe that, in experienced hands, echo-guided percutaneous BCDL cannulation, like in the adult population, could be a valid alternative to surgical cutdown and semi-Seldinger technique, even in newborns and small children.

Back to Top | Article Outline


1. Rollins MD, Hubbard A, Zabrocki L, et al. Extracorporeal membrane oxygenation cannulation trends for pediatric respiratory failure and central nervous system injury. J Pediatr Surg 2012; 47:68–75
2. Skinner SC, Hirschl RB, Bartlett RH. Extracorporeal life support. Semin Pediatr Surg 2006; 15:242–250
3. Lorusso R, Taccone FS, Belliato M, et al; Euro-ELSO Working Group on Neurologic Monitoring and Outcome: Brain monitoring in adult and pediatric ECMO patients: The importance of early and late assessments. Minerva Anestesiol 2017; 83:1061–1074
4. Polito A, Barrett CS, Wypij D, et al. Neurologic complications in neonates supported with extracorporeal membrane oxygenation. An analysis of ELSO registry data. Intensive Care Med 2013; 39:1594–1601
5. Cairo SB, Arbuthnot M, Boomer L, et al; American Pediatric Surgical Association Critical Care Committee: Comparing percutaneous to open access for extracorporeal membrane oxygenation in pediatric respiratory failure. Pediatr Crit Care Med 2018; 19:981–991
6. Rycus P. ECLS Registry Report, International Summary, July 2016. Ann Arbor, MI, Extracorporeal Life Support Organization, 1–26
7. Javidfar J, Brodie D, Wang D, et al. Use of bicaval dual-lumen catheter for adult venovenous extracorporeal membrane oxygenation. Ann Thorac Surg 2011; 91:1763–1768; discussion 1769
8. Bermudez CA, Rocha RV, Sappington PL, et al. Initial experience with single cannulation for venovenous extracorporeal oxygenation in adults. Ann Thorac Surg 2010; 90:991–995
9. Lazar DA, Cass DL, Olutoye OO, et al. Venovenous cannulation for extracorporeal membrane oxygenation using a bicaval dual-lumen catheter in neonates. J Pediatr Surg 2012; 47:430–434
10. Fallon SC, Shekerdemian LS, Olutoye OO, et al. Initial experience with single-vessel cannulation for venovenous extracorporeal membrane oxygenation in pediatric respiratory failure. Pediatr Crit Care Med 2013; 14:366–373
11. Rycus P. ECLS Registry Report, International Summary, January 2018. Ann Arbor, MI, Extracorporeal Life Support Organization, 1–37
12. Moscatelli A, Buratti S, Gregoretti C, et al. Emergency percutaneous, bicaval double-lumen, ECMO cannulation in neonates and infants: Insights from three consecutive cases. Int J Artif Organs 2015; 38:517–521
13. Speggiorin S, Robinson SG, Harvey C, et al. Experience with the Avalon® bicaval double-lumen veno-venous cannula for neonatal respiratory ECMO. Perfusion 2015; 30:250–254
14. Jarboe MD, Gadepalli SK, Church JT, et al. Avalon catheters in pediatric patients requiring ECMO: Placement and migration problems. J Pediatr Surg 2017; 53:1–4
15. Burns J, Cooper E, Salt G, et al. Retrospective observational review of percutaneous cannulation for extracorporeal membrane oxygenation. ASAIO J 2016; 62:325–328
16. Giordano P, Saracco P, Grassi M, et al; Italian Association of Pediatric Hematology and Oncology (AIEOP): Recommendations for the use of long-term central venous catheter (CVC) in children with hemato-oncological disorders: Management of CVC-related occlusion and CVC-related thrombosis. On behalf of the coagulation defects working group and the supportive therapy working group of the Italian Association of Pediatric Hematology and Oncology (AIEOP). Ann Hematol 2015; 94:1765–1776
17. Yastrebov K, Manganas C, Kapalli T, et al. Right ventricular loop indicating malposition of J-wire introducer for double lumen bicaval venovenous extracorporeal membrane oxygenation (VV ECMO) cannula. Heart Lung Circ 2014; 23:e4–e7
18. Moscatelli A, Pezzato S, Lista G, et al. Venovenous ECMO for congenital diaphragmatic hernia: Role of ductal patency and lung recruitment. Pediatrics 2016; 138:e20161034
19. Bartlett RH, Zwischenberger JB. Annich GM, Lynch WR, MacLaren G, et al; The circuit. In: ECMO, Extracorporeal Cardiopulmonary Support in Critical Care. 2012, pp Fourth Edition. Ann Arbor, MI, Extracorporeal Life Support Organization, 107–132
20. Toomasian JM, Lawson DS, Harris WE. Annich GM, Lynch WR, MacLaren G, et al; Management of blood flow and gas exchange during ECLS. In: ECMO, Extracorporeal Cardiopulmonary Support in Critical Care. 2012, pp Fourth Edition. Ann Arbor, MI, Extracorporeal Life Support Organization, 149–156
21. Pranikoff T. Brogan TV, Lequier L, Lorusso R, et al; Weaning and decannulation of children with respiratory failure on ECLS. In: Extracorporeal Life Support: The ELSO Red Book. 2017, pp Fifth Edition. Ann Arbor, MI, Extracorporeal Life Support Organization, 291–296
    22. Mazzeffi M, Fischer S. Transesophageal echocardiographic guided placement of a right internal jugular dual-lumen venovenous extracorporeal membrane oxygenation (ECMO) catheter. J Cardiothorac Vasc Anesth 2013; 27:e46–e47
    23. Gow KW, Heiss KF, Wulkan ML, et al; Extracorporeal life support for support of children with malignancy and respiratory or cardiac failure: The extracorporeal life support experience. Crit Care Med 2009; 37:1308–1316
    24. Gow KW, Wulkan ML, Heiss KF, et al; Extracorporeal membrane oxygenation for support of children after hematopoietic stem cell transplantation: The Extracorporeal Life Support Organization experience. J Pediatr Surg 2006; 41:662–667

    Avalon; extracorporeal membrane oxygenation; infants; newborns; percutaneous

    Supplemental Digital Content

    Back to Top | Article Outline
    ©2019The Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies