Right Atrial Lines as Primary Access for Postoperative Pediatric Cardiac Patients

To characterize the use of right atrial lines (RALs) as primary access in the postoperative care of neonatal and pediatric patients after cardiothoracic surgery and to identify risk factors associated with RAL complications. Observational retrospective cohort study in pediatric cardiac patients who underwent RAL placement in a tertiary children’s hospital from January 2011 through June 2018. A total of 692 children with congenital heart disease underwent 815 RAL placements during the same or subsequent cardiothoracic surgeries during the study period. Median age and weight were 22 days (IQR 7–134) and 3.6 kg (IQR 3.1–5.3), respectively. Neonates accounted for 53.5% of patients and those with single-ventricle physiology were 35.4%. Palliation surgery (shunts, cavo-pulmonary connections, hybrid procedures, and pulmonary artery bandings) accounted for 38%. Survival to hospital discharge was 95.5%. Median RAL duration was 11 days (IQR 7–19) with a median RAL removal to hospital discharge time of 0 days (IQR 0–3). Thrombosis and migration were the most prevalent complications (1.7% each), followed by malfunction (1.4%) and infection (0.7%). Adverse events associated with complications were seen in 12 (1.4%) of these RAL placements: decrease in hemoglobin (n = 1), tamponade requiring pericardiocentesis (n = 3), pleural effusion requiring chest tube (n = 2), and need for antimicrobials (n = 6). Multivariable logistic regression showed that RAL duration (OR 1.01, p = 0.006) and palliation surgery (OR 2.38, p = 0.015) were significant and independent factors for complications. The use of RALs as primary access in postoperative pediatric cardiac patients seems to be feasible and safe. Our overall incidence of complications from prolonged use of RALs remained similar or lower to that reported with short-term use of these lines. While RAL duration and palliation surgeries seemed to be associated with complications, severity of illness could be a confounding factor. A prospective assessment of RAL complications may improve outcomes in this medically complex population.


Introduction
The postoperative management of neonatal and pediatric patients undergoing cardiac surgery has evolved considerably over the past decades [1]. The extensive range of surgical options along with the increased patient complexity requires precise postoperative monitoring, as well as secure reliable access to provide care [1]. Right atrial lines (RALs) are widely used in most institutions for limited periods of time as a routine part of the postoperative care to monitor central venous pressures and right ventricular compliance and to administer medications and parenteral nutrition [2,3]. They are also indicated when percutaneous central venous access has failed, is difficult or technically challenging (previous thrombosis and congenital stenosis), or there is risk of compromising vessel patency (small neonates) [2,3]. In patients with single-ventricle physiology, these lines play a vital role in vascular preservation of vessels that are a part of future palliative pathways [2,3]. Some studies suggest that the use of RALs can decrease the incidence of venous thromboembolism, vessel occlusion, limb ischemia, chylothorax, and infection associated with percutaneous central catheters [3]. Peripheral vein health may also be preserved with the use of these lines. Despite the widespread use of RALs, there are limited data on complications and factors associated with their occurrence [2][3][4][5][6][7][8][9][10][11]. Additionally, guidelines for standard insertion, management, and removal of transthoracic lines (TTLs) are lacking. The aim of our study was to characterize the use of RALs as prolonged primary access in the postoperative care of neonatal and pediatric cardiac patients and to identify factors associated with RAL complications in this population.

Study Setting and Design
The study was conducted in a tertiary-level multidisciplinary cardiovascular intensive care unit (CVICU). It was designed as an observational, retrospective cohort study. Approval from the Institutional Review Board was obtained, with waiver of informed consent granted prior to study initiation.

Study Population and Data Collection
All pediatric cardiac patients 1 day to 18 years of age who underwent RAL placement between January 2011 and June 2018 were included in the study. Patients with incomplete or inaccurate data were excluded from the study.
Database included patient characteristics, cardiac and surgical data, RAL characteristics, presence and duration of other central lines, laboratory values, complications, and outcomes. Patient characteristics included age, weight, gender, race, prematurity information, prenatal cardiac diagnoses, non-cardiac abnormalities, and presence of a syndrome or association as well as genetic diagnoses. Cardiac diagnoses were dichotomized as single-and non-single-ventricle physiology. Cardiothoracic surgeries were divided into palliation (shunt, Glenn, Fontan, hybrid procedure, and pulmonary artery banding) and non-palliation. Cardiopulmonary bypass (CPB), surgical times, and the need for preoperative and postoperative mechanical circulatory support were included.
RAL characteristics included number of lumens, line duration, and time from RAL removal to hospital discharge. RAL complications were divided into thrombosis, migration, malfunction, infection, bleeding at removal, accidental removal, and line retention. Complications were recorded via chart review, including review of daily notes, echocardiogram reports, blood cultures, and nursing safety reports. Thrombosis was defined as inability to aspirate or infuse through the RAL requiring thrombolysis or the presence of a line-related thrombus in the right atrium on echocardiogram. Migration was defined as inadvertent extrusion of the catheter outside of the right atrium. Malfunction was defined as dampened or no waveform tracing, the presence of a leakage or catheter rupture, or the inability to infuse through or aspirate blood from the RAL not related to the presence of a thrombus. Infection was defined as positive blood culture obtained from the RAL with clinical signs of infection. Bleeding at line removal was considered present when it led to hemodynamic compromise, when it required an intervention (such as fluid resuscitation or the need for a transfusion), or when a pericardial effusion was found in the echocardiogram post RAL removal. Patients with increased chest tube output not fulfilling the prior definitions were not considered as having this complication. Accidental removal was defined when the RAL was removed accidentally prior to the planned date. Line retention was defined when regular removal failed requiring a surgical intervention or procedure. Laboratory values (hemoglobin, platelet count, and coagulation profile) as well as RAL pressure at the time of removal were also collected. Outcomes included survival to hospital discharge, length of hospital stay (LOS), surgery to discharge time, mechanical ventilation (MV) duration, and failed extubations.

Intraoperative and Postoperative Central Line Management
Our usual practice for patients with unreliable, insufficient, or no central access prior to surgery is to have a central line placed (internal jugular, femoral, or subclavian) by anesthesia in the operating room (OR) for case management during the surgery. At the end of the case, RAL is placed by the cardiothoracic surgeon and checked for adequate functioning as explained below. All continuous medications are moved to the RAL within the first postoperative day if patient is stable and the RAL is used a primary access point throughout the postoperative course. The other central lines are usually removed within 72 h after the surgery unless additional central access is required or if the RAL cannot be safely used due to the presence of a complication.

RAL Characteristics and Placement
Our placement technique is tailored to the individual patient and starts with preoperative assessment. In general, we use double lumen catheters of two different types. A silastic umbilical catheter is preferred for shunt-dependent patients or those with vulnerable circulations at the time of catheter removal (to prevent significant catecholamine surges). We usually place a cuffed catheter to the remaining patients due to the increased security and more difficulty to migrate (which usually requires more resources as well as sedation at line removal). At the end of CPB or before chest closure for non-CPB cases, a pledgeted (with autologous pericardium) suture is placed circumferentially around the atrial appendage. A "jake" clamp is passed subcutaneously from the median sternotomy wound through the subcutaneous tissues to exit the skin in the infraclavicular region at the clavipectoral junction (using this clamp the catheter is pulled into the mediastinal space). The catheter is then connected to an extension tubing and then to a three-way stopcock so that line changes can be made up to the extension tubing (since it is easier to replace the extension tubing if there is damage during changeover versus the connectors on the actual catheter). For cuffed catheters, we believe that it is necessary to cut the catheter to the required length. In general, we mark the required catheter in 1-cm increments starting from the cut end with 3 marks (one at 1 cm from the end, another at 2 cm from the end, and a third at 3 cm from the end). The umbilical catheters are generally already marked, and we tend not to cut them to size. An incision is made in the atrial appendage and the catheter tip is introduced directly into the atrium usually to the third mark. The pledgeted suture is tied carefully around the catheter entrance site. The purse string is tied tight enough to prevent leaking around the catheter but avoiding deformation of the catheter. A mini silastic vessel loop or a lacrimal duct tubing is then wrapped around the atrial appendage and catheter, thus applying pressure but not deforming the catheter. This silastic tubing is tied in 2 knot throughs and 2 small hema-clips are placed on the ends of this tubing. The silastic tubing is then cut flush with the outermost clip. The line is then tested for ability to draw and flush easily. Prior to sternal closure the line is pulled back to the second mark, so that the catheter is 2 cm into the atrium from the entrance site. Once the sternum is closed, the line is then secured in a loop in the left infraclavicular/pectoral region with 4 silk sutures. The umbilical catheter is coiled upon itself so that the end of the catheter is adjacent to the shoulder, and the loops are secured with the silk sutures. For the cuffed catheter, the cuff is placed just under the skin. The length between the cuff and the union of the 2 separate infusion ports is coiled and secured in a loop on the chest wall. The actual union site is then secured to the anterior shoulder with 2 sutures which are first tied to the skin and then brought through the union piece and secured down by tying the suture securely on this union. The catheter is positioned so that the dressing does not overlap the sternal wound dressing for independent exchange (Fig. 1).

RAL Management and Removal
RAL position is confirmed with chest X-ray (CXR) on arrival to the CVICU. For line patency, heparinized fluids (0.5 units of heparin/ml) are infused continuously through the line. Bedside nurses check for line functioning as well as patency of dressing and suture at least once per shift, and daily CXRs are performed and evaluated to assess for RAL position. On daily CXR, the catheter tip is measured from the 2 hema-clips placed at the entrance site, with the "tip to clip" distance being ideally 1 to 2 cm. We believe that if this distance is greater than 2 cm, the catheter can obstruct from pressing against the atrial wall, which can also promote arrhythmia. If less than 1 cm, the catheter can be at risk of migration into the pericardium. If migration is suspected, an echocardiogram is performed. Platelet count and coagulation labs are usually sent prior to planned removal. Lines are removed only by the cardiothoracic surgical team at the bedside unless the patient requires another procedure to be performed in the OR. Catheter removal involves appropriate sedation, removal of sutures, and careful traction. For the cuffed catheters that have been in over 3 weeks, the tissue at the entrance site is carefully dissected, the cuff is grasped, and traction is applied to remove the catheter. Bleeding is generally limited, and the exit site is typically dressed with a sterile dressing. If bleeding occurs, direct pressure is applied until bleeding stops, and the patient is placed in anti-Trendelenburg position to decrease central venous pressure. An echocardiogram is performed at 4-h post-removal.

Statistical Analysis
Demographic characteristics, cardiac and surgical data, RAL characteristics, laboratory data, and complications as well as outcomes were described using medians and interquartile ranges (IQR) for continuous variables and frequencies and percentages for categorical variables. Bivariate analyses were conducted using χ 2 tests and Wilcoxon-Mann-Whitney tests to ascertain the association between covariates and RAL complications. Multivariable logistic regression models were used to analyze the effects of potential variables on the probability of complications. Backward selection with an alpha level removal of 0.05 was utilized. Odds ratios (OR) and 95% confidence intervals were calculated. All p-values were 2-sided and p < 0.05 was considered statistically significant. Statistical analyses were performed using SAS (version 9.4, SAS Institute Inc., NC, USA).

Patient Population
A total of 692 neonates and children with congenital heart disease underwent 815 RAL placements between January 2011 and June 2018. Median age and weight of the cohort were 22 days (IQR 7-134) and 3.6 kg (IQR 3.1-5.3). Neonates (< 30 days old) accounted for 53.5% of patients. Males were predominant (55.5%). One-third of the patients had a clinical diagnosis of a syndrome or presumed genetic diagnosis. Cardiac diagnoses with single-ventricle physiology accounted for 35.4%. Of the 815 cardiothoracic surgeries, 82.7% required CPB and 38% were palliations (Blalock-Taussig-Thomas and central shunts, cavo-pulmonary connections, hybrid procedures, and pulmonary artery bandings). Survival to hospital discharge reached 95.5%. Table 1 summarizes the population and line characteristics among the 692 unique patients.

RAL and Other Central Line Characteristics
Most patients had only 1 RAL in place (85%), while the others had multiple RALs placed either at the primary or subsequent surgeries during the same hospitalization (range 2-3 RALs). Double lumen RALs were the most prevalent (66.8%). Median RAL duration was 11 days (IQR 7-19) with a range of 1 to 399 days. The median RAL removal to hospital discharge time was 0 days (IQR 0-3), with 61% of all patients having their RAL removed on the day of hospital discharge. Six patients were discharged home with a RAL in place for treatment continuation.

Total RALs and Complications
A total of 46 complications occurred among the 815 RALs (5.6%), with one patient having 2 RAL complications during different postoperative periods. Among the whole cohort of patients, those with complications (n = 45) had a median age of 47 days (IQR 9-154) and 44% of them were neonates. Prematurity accounted for 20%. Non-cardiac abnormalities, clinical syndrome/association, and genetic diagnosis were found in 28.9%, 33%, and 20% of them, respectively. Twenty (40%) patients had single-ventricle physiology that required a shunt (n = 12), pulmonary artery banding (n = 3), hybrid procedure (n = 2), Glenn (n = 2), and Fontan (n = 1). Some patients without single-ventricle physiology but with complex anatomy also required palliation surgeries as first step prior to biventricular repair. Palliation surgeries overall accounted for 53% of the cohort of patients with complications. Survival to hospital discharge reached 88.9% in this group.
Of the total number of RALs (n = 815), thrombosis and migration were the most prevalent complications (1.7% each), followed by malfunction (1.4%) and infection (0.7%). When RALs with and without complications were compared, no baseline differences were observed in patient characteristics, cardiac and surgical data, survival to hospital discharge, or laboratory values and RAL pressures at line removal. RALs with complications had significantly longer duration (19 days vs. 11 days, p = 0.0008) and worse patient outcomes: longer LOS (48 days vs. 20 days, p < 0.0001), more prolonged surgery to discharge time (27 days vs. 13 days, p < 0.0001), longer MV duration (9 days vs. 3 days, p = 0.001), and more failed extubations (45.7% vs. 24.7%, p = 0.001). RAL removal to hospital discharge time was also longer in RALs with complications (4 days vs. 0 days, p < 0.0001). Table 2 compares characteristics of RALs with and without complications.

RAL Complications and Factors
Thrombosis occurred in 14 RALs (1.7%), requiring thrombolysis or line removal. No effects to patients were found due to this complication except for a decrease in hemoglobin due to thrombi-related leakage in 1 patient requiring blood transfusion. Migration occurred in 14 RALs (1.7%), with 9 of those (64%) migrating to the pericardial space and 5 (36%) to the pleural space. Line removal was required in all cases. Of those, 3 patients had tamponade physiology requiring emergent pericardial drainage and 2 required chest tube placement. There were Medians (IQR) are used to describe continuous variables and frequencies (%) are used for categorical variables ECMO extracorporeal membrane oxygenation, g/dL grams per deciliter, INR international normalized ratio, mg/dL milligrams per deciliter, μL microliter, mmHg millimeter of mercury, PA pulmonary artery, PICC peripherally inserted central catheter, PTT partial thromboplastin time, RAL right atrium line, s seconds, UVC umbilical venous catheter, VAD ventricular assist device   7 (4-10) 6 (4-10) 7 (4-10) 0.373

Discussion
This study is the first to date to uniquely characterize the prolonged use of RALs as primary access in postoperative neonatal and pediatric cardiac patients and to identify factors associated with RAL complications in this population. Our primary finding is that the incidence of RAL complications in our cohort was low. Compared to previous TTL studies, our rates of thrombosis and migration were similar while those of malfunction, infection, bleeding at removal, accidental removal, and line retention were lower in our patient population [2,11]. Similar to previous TTL studies that reported rates of thrombosis and migration ranging 0.4-4% [3,5,6,[9][10][11] [3,8], and 0.07-0.7% [3,4,6], respectively. These later findings could be related to our insertion and line securement techniques.
Interpreting TTL complication literature is challenging due to multiple factors. First, existing studies are very limited, and some were published decades ago, which challenges any longitudinal assessment of the data. Second, all the studies are retrospective reviews from single centers with different patient populations and study designs. Third, there are multiple discrepancies or lack in their complication definitions. For instance, bleeding at the time of TTL removal was defined varyingly across studies as increase in chest tube output, the need for a blood transfusion, hemodynamic compromise, the need for an intervention, and the presence of a pericardial effusion on echocardiogram or tamponade physiology, which may account in part for the variability in the incidence of this specific complication. Furthermore, TTL are center specific with different indications, insertion techniques, monitoring, and protocols for removal, making interpretation of the current literature more conflicting. Table 5 summarizes the literature review [2][3][4][5][6][8][9][10][11].
Another factor to consider when interpreting complications is line duration. While historical TTL reports failed to provide clear data on line duration [4][5][6], our study reported longer line duration than most recent reports [2,3,8,10,11]. Furthermore, 61% of our patients had their RAL removed at hospital discharge and 6 patients went home with the RAL in place for continuation of therapies. No increase in complications were observed with any of these practices in our patient population. This factor would be expected to portend worse outcomes to our patients since previous pediatric cardiac surgical studies like ours demonstrated that line duration (for both central venous lines and TTLs) is a significant predictor for line complications [5,[11][12][13][14]. Severity of illness could be an important confounding factor in this association since we tend to maintain central access longer in sicker patients and in those at risk for hemodynamic decompensation. Furthermore, the upper limit for line duration resulting in RAL complications continues to remain unknown and warrants more investigation.
We observed that palliation surgery was a significant factor for RAL complications in our cohort. As compared with other cardiac surgical patients, patients undergoing multistage single-ventricle palliation have significant morbidity, extracardiac abnormalities, and specifically anatomic complexities (such as heterotaxy syndrome) that could explain more RAL complications [15,16]. They also tend to have increased severity of illness, leading to requirement of stable access for longer periods of time, which may also contribute to increased rates of complications [15][16][17].
Over the last few decades, TTLs have played an increasingly important role in the postoperative management of pediatric cardiac surgical patients. While other centers use central venous lines and PICCs as fundamental access sites, we report the use of RALs as primary (and sometimes unique) access for the postoperative management of neonatal and pediatric cardiac patients. RALs seem to be a safe and essential tool with several undefined advantages. Vascular preservation of important vessels, particularly those utilized in single-ventricle palliation pathways, is an important advantage to these lines. In a multi-institutional cohort study utilizing the Pediatric Cardiac Critical Care Consortium (PC4) registry, 31% of postcardiac surgical hospitalizations required a TTL. They reached 84% after Norwood operation and 56% after cavo-pulmonary connections [11]. Vessel patency ensures ongoing candidacy for palliative surgeries or transplant [17]. Access point reliability in extremely complicated and medically fragile patients is another benefit to these lines; particularly in the single-ventricle population, who often have long postoperative courses with a high burden of morbidity requiring stable and durable access. The health of peripheral veins is also important for lifetime care in this and other cardiac patient populations. Furthermore, this access reliability minimizes morbidity and risks associated with subsequent anesthetic courses required to gain stable access when needed in these patients [18][19][20].
Our study has a few limitations. First, it is a retrospective database review of a single-center experience with limited data. Second, we were not able to adjust for presumed factors associated with RAL complications, such as severity of illness. Finally, a learning curve effect due to institutional