Major Short-term Complications of Arterial Cannulation for Monitoring in Children

This study was deemed a minimal risk and the requirement to obtain informed consent was waived in accordance with 45 CFR 46.116, as approved by the Mayo Clinic Institutional Review Board (IRB #15-000407 Rochester, Minnesota). We performed a single-center, retrospective, cohort study. Our methodology was similar to a previous study from our institution that examined brachial artery cannulation complications in adults.¹¹ 

We identified all pediatric (age less than 18 yr) surgical patients who underwent arterial cannulation for monitoring purposes from January 1, 2006 to August 15, 2016 from an institutional database that contains information on all patients who have received anesthetic care at Mayo Clinic (Rochester, Minnesota) from 1998 to present. Patients were excluded if they (1) were 18 yr and older; (2) revoked the right to use their medical records for research; (3) had arterial catheterization performed for diagnostic purposes (such as in the cardiac catheterization laboratory); or (4) underwent arterial line placement outside of the perioperative setting (such as in the intensive care unit).

Data were collected from three primary sources: (1) the Mayo Clinic Perioperative DataMart, (2) Mayo Clinic Life Science Services and the Data Discovery Query Builder, and (3) the Mayo Clinic Infection Prevention and Control database. The Perioperative DataMart is the access layer of an institutional data warehouse containing detailed information regarding aspects of a patient’s surgical encounter (e.g., demographic information; procedural descriptions; locations; start and stop times; detailed physiologic information including vital signs, ventilator data, laboratory information; and fluid, transfusion, and medication administration information).¹²  This database also contains detailed information regarding all anesthesia-related procedures, including invasive catheter placements. Pertinent data include the anatomic location of cannulation, size of the catheter placed, and procedure-related complications. In addition to these catheter-related details, specific variables extracted from the Perioperative DataMart included the following: height, weight, age, date of birth, and sex. All details related to the cannulation procedure, surgical procedure, anesthesia start and stop times, surgery start and stop times, and American Society of Anesthesiologists physical status score were also collected from the Perioperative DataMart, as was the need for a second operation within 30 days of catheter placement.

Our second primary data source (Mayo Clinic Life Science Services and the Data Discovery Query Builder) has been described and accuracy demonstrated.^13,14  In brief, Mayo Clinic Life Science Services is a near real-time replicate of the institution’s electronic medical record system. This data warehouse is developed from multiple original clinical data sources, including highly annotated, full-text clinical notes, laboratory tests, diagnostic findings, demographics, and related clinical data from the year 2000 onward. The Mayo Clinic Life Science Services provides a query-building tool, the Data Discovery Query Builder, a Web-based application that is intended to help researchers interrogate data files contained in Mayo Clinic Life Science Services. The Data Discovery Query Builder allows users to identify administrative, demographic, laboratory, and diagnostic data of interest within the electronic medical record. Mayo Clinic Life Science Services and the Data Discovery Query Builder were also used to identify our outcomes of interest.

Our third data source was the Infection Prevention and Control Committee at the Mayo Clinic (Rochester, MN), described further in the “Infectious Complications” section.

Complications of arterial cannulation were placed into three categories: vascular, neurologic, or infectious.

To identify vascular complications, all patients who had a vascular medicine, vascular surgery, hematology, or orthopedic surgery consultation in the 30-day interval after catheter placement were identified. The indication for the consultation was reviewed to evaluate whether the consultation may have been related to a catheter-related complication. Key terms queried included limb or digital ischemia, thrombosis, compartment syndrome, hematoma, pseudoaneurysm, discoloration, and cool extremity.

To further enhance our catchment of vascular complications, all patients who underwent vascular ultrasound diagnostic evaluation or who returned to the operating room within 30 days of catheter placement were evaluated to determine: (1) whether the subsequent ultrasound scan or procedure was performed on the extremity in which arterial cannulation had been performed and if so, (2) whether the ultrasound scan report revealed a cannulation-related complication (such as thrombosis) or the procedure was required to address a cannulation-related complication.

Similar procedures were performed for neurologic complications, including identification of patients who had a neurologic, neurologic surgery, or orthopedic surgery consultation. Query terms of interest included burning, paresthesia, palsy, weakness, loss of sensation, tingling, or neuropathy.

Cases of catheter-related bloodstream infection were identified using the resources of Infection Prevention and Control Committee based on Centers for Disease Control and Prevention guidelines.¹⁵  The Infection Prevention and Control Committee at our institution conducts surveillance of patients with positive blood cultures who have undergone invasive procedures, including arterial catheter placement. If a case of suspected catheter-related bloodstream infection is identified with these surveillance methods, an infection control practitioner reviews the patient’s medical record and determines whether the infection was related to an arterial catheter placement. For this study, documentation of catheter-related bloodstream infection by the Infection Prevention and Control Committee was required within 30 days of catheter placement.

All patients identified as having had a potential catheter-related complication as previously defined had their electronic medical record examined by independent reviewers (A.V.W., K.S.H., D.E.T.). Each individual reviewed the records independently to determine whether the complication was associated with the arterial line. All complications were reviewed by two senior reviewers (S.J.G., M.E.N.) to confirm associated complications.

In order to enhance the validity of our query methodology, we evaluated 100 random patients who underwent arterial cannulation, but who were not identified from our defined search strategies to detect possible complications. Among these 100 random “non-hits,” none were found to have any associated complication related to arterial cannulation upon manual abstraction of their records.

We performed a descriptive analysis of the retrospective cohort. A data analysis and statistical plan was written after the data were accessed. For summary statistics, dichotomous variables are presented as numbers and percentages while continuous data are presented as medians with 25 to 75% interquartile ranges. For the purposes of analysis, separate arterial line cannulation events within the same participant were treated as independent events (i.e., these were not treated as clustered or correlated). The Clopper–Pearson exact method was utilized to calculate binomial 95% CI for a binomial proportion when the data have already been observed and the observed number of complications in the retrospective cohort is zero.¹⁶  For excluded subjects, analyses were conducted assuming that these data were missing at random. All statistical analyses were performed using JMP statistical software Pro version 14.1.0 (SAS Institute, USA).

During the study period, there were a total of 111,934 surgical procedures in patients less than 18 yr of age, and 5,142 arterial cannulations were performed in 4,178 patients (fig. 1); 7,154 patients were excluded for denial of research authorization. Subject characteristics are shown in table 1. The overall median age was 9 yr (interquartile range, 2 to 14) and 51.4% were male. The most common anatomical sites for arterial cannulation were the radial (N = 3,395, 66.0%) and femoral arteries (N = 1,528, 29.7%). The ulnar, pedal, and brachial/axillary sites accounted for the remaining 216 cannulations (4.2%). There were three umbilical arterial cannulations performed for neonates delivered in the operating room with assistance from the neonatology resuscitation team. Figure 2 depicts the distribution of arterial cannulation site by age. The majority of arterial cannulations were in patients with American Society of Anesthesiologists Physical Status scores II and III (N = 4,062 [79%]).

Results of the chart review are shown in table 2. For potential vascular complications, 333 consultations from specialty services, 150 return-to-the-operating-room events, and 28 diagnostic vascular ultrasounds were reviewed. From these, eight patients were determined to have vascular arterial cannulation-related complications. Query of our institutional database revealed three patients with catheter-related bloodstream infection attributed to the arterial line. Among 436 consultations and return-to-the-operating-room events, we did not find any short-term neurologic complications.

In total, there were 11 short-term major complications attributed to arterial cannulation (0.2%; overall major complication rate, 2 per 1,000 lines; 95% CI, 1 to 4). The major complication rates of arterial cannulation by anatomic site are shown in table 3. All 11 major complications occurred in femoral arterial lines (femoral anatomic site, 0.7%; major complication rate, 7 per 1,000 lines; 95% CI, 4 to 13).

Subject characteristics and age definitions of those who underwent femoral arterial cannulation are shown in table 4. The median age was 2 yr (interquartile range, 0.3 to 9). There were two complications in neonates (major complication rate, 11 per 1,000 lines; 95% CI, 3 to 39), six in infants (major complication rate, 16 per 1,000 lines; 95% CI, 7 to 34), and three in toddlers (major complication rate, 7 per 1,000 lines; 95% CI, 3 to 21). There were no complications in children older than 5 yr. The vast majority (90%) of femoral arterial lines were placed in children undergoing cardiac surgery, and 81.7% of patients were American Society of Anesthesiologists Physical Status III.

Patient and arterial cannulation details for each complication are described in table 5. Nine of the 11 patients had arterial cannulation for cardiac surgical procedures. Vascular complications occurred earlier in a patient’s course at median 1.5 days (interquartile range, 1 to 2.75) and ranged from loss of pulse requiring line removal and systemic anticoagulation to development of a large retroperitoneal hematoma from bleeding external iliac artery requiring surgical repair and hematoma evacuation. Catheter-related blood stream infections occurred later in the patient’s course at a median 11 days (interquartile range, 6 to 23). All infections were treated with removal of the arterial line and treatment with systemic antibiotic agents. Complications occurred throughout the study period (2007 to 2016). Ultrasound guidance for femoral arterial line placement was utilized in 9 of the 11 cases (1 missing data, 1 not used).

In this single-institution retrospective cohort of arterial cannulation for monitoring purposes in children, we describe the use patterns and rate of major complications. The overall major short-term complication rate was low (0.2%). All of the complications, including vascular complications and catheter-related blood stream infections, occurred with cannulation of the femoral artery in younger children younger than 5 yr old. There were no complications in cannulation of distal arteries, including 3,395 radial arterial lines.

Arterial cannulation in children in the perioperative setting is frequently performed for continuous blood pressure monitoring and blood sampling.^3,17  Cannulation of the radial artery is preferred for its anatomic accessibility, reliable presence of collateral blood supply, and low complication rate. However, such cannulation is not always successful, especially in neonates and infants.^3,18  The procedure is technically more challenging in children, particularly in younger and smaller children.^2,3  In addition, there are reported differences in blood pressure measurements in neonates and infants undergoing cardiac procedures.¹⁹  Consequently, alternative and proximal sites of arterial cannulation in neonates and infants may be sought, including femoral arterial cannulation, which is what we observed in our cohort.^5,6  Femoral arterial cannulation was the preferred site of arterial cannulation in the neonate and infant age groups.

Before this study, arterial cannulation of the femoral artery was the preferred method for monitoring in pediatric cardiac surgical procedures. As a result of our retrospective data and literature review, in late 2018, we changed our practice to preferentially use the radial or ulnar arteries in children less than 5 yr of age whenever possible. Cannulation of the femoral artery remains a backup site if the radial/ulnar sites are unsuccessful.

All of the short-term major complications related to arterial cannulation in children occurred in those with femoral arterial lines. The vast majority of studies examining complications of femoral arterial access originate from the pediatric cardiac catheterization literature in children with congenital heart disease.^10,20,21  Several small studies have examined complications related to femoral arterial cannulation for monitoring in the perioperative and intensive care unit environments.

In a study of 282 patients, DuMond et al. reported outcomes in younger patients less than 2 yr old who underwent femoral arterial cannulation for monitoring purposes in a cardiac surgical practice.⁶  Similar to our study, a greater degree of perfusion abnormalities, defined as pulse strength discrepancies, were found in younger patients with larger catheters (greater than 2.5 French), and in those with longer durations of indwelling arterial catheters. In addition, 20% of patients had documented pulse discrepancies and the incidence of the loss of distal pulse was 3.4%.

Similar complications were seen in a study of 77 femoral arterial cannulations in critically ill children.²²  There was an overall vascular complication rate of 11%; vascular complications were defined as decrease in distal pulses, pallor, and prolonged capillary refill. The infectious complication rate was 1.4%. We found a lower rate of major vascular complications compared to these two studies, which is likely due to methodologic differences, as we were not able to detect lesser complications related to arterial cannulations (e.g., pulse discrepancies or pulse loss) without accompanying diagnostic or therapeutic intervention.

In 2011, Brotschi et al. studied the incidence of arterial catheter-related thrombosis in critically ill children in a single-center multidisciplinary pediatric intensive care unit in Zurich.²³  Among 615 arterial lines included, 137 (22%) were placed in the femoral artery and 18 of those developed thrombus (13% incidence). Brotschi et al. used methodology similar to our study to identify arterial line thrombosis, including utilizing clinical signs of thrombosis. While their overall complication rate of femoral arterial lines was substantially greater than our study (13% vs. 0.7%), neither study found any complications in distal cannulation sites, including radial, ulnar, or pedal arterial lines.

Previous prospective data from our institution examining femoral arterial monitoring found diverse results.⁵  In data reported in 1987, 165 femoral arterial cannulations were studied in 151 patients, utilizing 18-e to 20-gauge catheters. Perfusion-related complications were defined as signs or symptoms of distal ischemia or decreased perfusion that led to removal of the arterial catheter. The overall complication rate was 2.4% for perfusion-related complications and 3.6% for infectious complications, higher than our reported rate of 0.2%. Neonates were again at the highest risk of perfusion-related complications (25%). This study also compared femoral arterial cannulation complication rates to previous reports of radial arterial cannulation. Surprisingly, in contrast to our study, vascular and infectious complications of femoral arterial lines were similar to two studies of radial arterial cannulation.^24,25  These differences may simply relate to the time periods studied. As noted, larger catheters were utilized during that time period. Additionally, greater focus on aseptic technique for all line placements in recent practice may also have led to a decreased infectious risk.

The age of the patient appears to be a consistent risk factor for femoral arterial line related complications. In the studies outlined here, younger children were at higher risk of complications.^5,6,22,23  Indeed, in our study, neonates and infants had the highest complication rates at 10.9 and 15.9 per 1,000 lines, respectively. The reason for this effect has not been fully elucidated, but may be due to the inability to immobilize the legs in young children, with resulting repeated flexion and extension movements causing damage to the vessel walls.²³  In addition, there may be catheter-related risk factors, such as the need for a longer catheter in the femoral location. Further study in this area is required. No major complications related to arterial catheters were noted in children age 6 yr or older, further highlighting the increased risk of femoral arterial lines in young children.

In general, direct comparisons between our studies and previous studies are difficult due to the wide variety of definitions utilized for complications, as well as differences in methodology. Larger, multicenter prospective studies with standardized identification and definition of complications are needed to more clearly identify risk factors and to provide better guidance for clinicians to minimize risks of arterial cannulation in children.⁸ 

This study has several limitations. Our study spanned a 10-yr period and consequently, may be confounded by factors that changed over time. There were 7,154 patients who declined research authorization, which may impart selection bias in our analyzed cohort. Approximately 30% of arterial lines were placed in the femoral artery, which may be higher than other pediatric practices and may inhibit broad applicability of our study. While similar methodology has been used previously for related studies, we recognize limitations due to its retrospective nature and reliance on information that exists at the time of data collection. We relied on database queries to identify complications and it is possible that some major complications were missed. Due to a major change in our institution’s electronic medical record in early 2018, our data is limited to the current study period utilizing our validated methodology. We cannot exclude other factors that may have contributed to the difference in complications observed, including catheter size to vessel–diameter ratio. In addition, we were not able to reliably determine the duration of arterial cannulation, which prohibited further temporal analysis. Because our overall major complication rate was low, multivariate analysis was not possible.

In conclusion, the overall major complication rate of arterial cannulation for monitoring purposes in children is low (0.2%). However, all major complications were observed in arterial lines in the femoral anatomic location in children less than 5 yr of age, with the greatest complication rates in neonates and infants. There were no complications in cannulation of distal arterial sites, including more than 3,000 radial arterial cannulations.

Support was provided solely from institutional and/or departmental sources.

The authors declare no competing interests.