Chapter 12. Intravascular Devices

• The subclavian approach is preferred for placement of central venous catheters (CVCs).
• Real-time ultrasound may reduce the mechanical complications associated with CVC insertion.
• Chlorhexidine-based skin antiseptic solutions reduce the incidence of catheter-related bloodstream infections as compared with povidone-iodine.
• Almost 50% of hospital-acquired bloodstream infections are caused by staphylococcal species.
• CVCs should not be replaced nor exchanged over a guidewire on a routine basis.

Central venous catheters (CVCs) have become an integral part of delivering care in the modern ICU. In fact, the Centers for Disease Control and Prevention (CDC) estimates that in U.S. ICUs there are 15 million CVC days per year (total number of days patients are exposed to CVCs).1 Indications for placement of CVCs include invasive hemodynamic monitoring, administration of vasoactive drugs, administration of caustic agents (e.g., chemotherapy), administration of parental nutrition, renal replacement therapy, large-bore venous access for rapid administration of fluids, and long-term venous access. This chapter focuses on the use of CVCs in the ICU setting. Thus long-term tunneled catheters used for hemodialysis and peripherally inserted central catheters (PICCs) will not be discussed.

The clinical presentation often dictates the type of catheter to be inserted. For example, a patient with a hemodynamically significant gastrointestinal hemorrhage may require only a single-lumen, large-bore CVC for volume resuscitation in addition to a peripheral IV, whereas a neutropenic patient with septic shock may require a triple-lumen CVC in order to administer vasoactive drugs and antibiotics simultaneously. Importantly, most evidence suggests that the number of catheter lumens does not affect the rate of CVC infectious complications.2,3 Once the type of catheter has been selected, an anatomic site for insertion needs to be determined. The optimal anatomic location for insertion of CVCs has been a matter of debate for many years. In 2001, Merrer and colleagues4 published a study of 289 patients who were randomized to have their CVCs inserted in either the femoral or subclavian vein. Patients with femoral vein catheters had a dramatically higher incidence of infectious complications (19.8% versus 4.5%; p <0.001) and thrombotic complications (21.5% versus 1.9%; p <0.001) as compared with patients with subclavian catheters. The overall sum of mechanical complications (i.e., arterial puncture, pneumothorax, hematoma or bleeding, and air embolism) was similar between the two groups. To date, there are no randomized trials comparing subclavian versus internal jugular catheters with regard to infectious complications, although observational studies suggest a lower rate of infectious complications with subclavian catheters and a similar rate of mechanical complications.5,6 As a result of these and other studies,7 the CDC recommends that, if not contraindicated, the subclavian vein should be used for the insertion of nontunneled CVCs in adult patients in an effort to minimize infection risk.

Infraclavicular Subclavian Approach

Prior to the insertion of an infraclavicular subclavian CVC, a small rolled-up towel should be placed between the shoulder blades to move the vascular structures to a more anterior position. After the subclavian area has been sterilely prepped and draped (see below) and local anesthesia has been administered, the patient should be placed in the Trendelenburg position. The arm should be positioned at the patient's side so that the shoulder, clavicle, and sternal notch are aligned and perpendicular to the sternum. The subclavian vein arises from the axillary vein and travels beneath the clavicle and inferior to the subclavian artery prior to joining the internal jugular vein and forming the brachiocephalic vein. Thus the clavicle provides a good anatomic landmark for the insertion of a subclavian CVC. The skin and periosteum of the clavicle should be infiltrated with a local anesthetic (e.g., lidocaine 1%). Following administration of local anesthesia, the skin should be entered with a 16-gauge introducer needle 1 to 2.5 cm below the inferior edge of the clavicle and 2 to 4 cm lateral to the midpoint of the clavicle. Once the needle is directly underneath the clavicle, it should be advanced toward the sternal notch, making sure that the needle remains in the plane immediately below the clavicle. If no blood return is obtained, then the needle should be pulled back and directed more cephalad. Slight backpressure should be placed on the plunger of the syringe any time the needle is advanced or withdrawn so that blood return can be visualized when the vessel is cannulated. After the vessel has been accessed, the modified Seldinger technique is used to complete the insertion of the CVC. Inexperienced operators often have difficulty knowing when the needle is directly underneath the clavicle and are appropriately fearful of puncturing the visceral pleura. Thus we often use a slightly different approach when supervising an inexperienced operator. Once the skin is entered, the operator is instructed to find the clavicle with the tip of the introducer needle. After the edge of the clavicle is reached and more local anesthesia is given in the area, the introducer needle is removed slightly (1 cm) and redirected in a more posterior direction by pushing down on the syringe and needle as a unit with the nondominant hand. This prevents the inexperienced operator from advancing the needle at a steep angle toward the underlying visceral pleura of the lung while attempting to locate the posterior border of the clavicle. Once the needle is “walked down” the bone and is positioned underneath the clavicle, it is then redirected toward the sternal notch and advanced slowly while applying backpressure to the syringe (Fig. 12-1).

Internal Jugular Approach

If the patient's anatomy (e.g., scar from previous vascular access) or a coagulation disorder prevents use of the subclavian vein, then the internal jugular vein should be used for placement of a CVC. The central approach to placing an internal jugular catheter uses the triangle formed by the two heads of the sternocleidomastoid muscle and the medial portion of the clavicle as the anatomic landmark. Typically, the internal jugular vein is lateral to the carotid artery, and both vessels run through the triangle beneath the sternocleidomastoid muscle. After the patient has been sterilely prepped and draped and local anesthesia has been administered, the patient is placed in the Trendelenburg position, and the head is rotated slightly toward the contralateral side such that carotid artery can be palpated in the apex of the triangle. The nondominant hand is used to lightly palpate the carotid artery, with careful attention paid to not place too much pressure on the skin because this can alter the position of the internal jugular vein. An 18- or 20-gauge “finder needle” often is used to locate the vessel prior to using the introducer needle (16 gauge). This finder needle should enter the skin at the apex of the triangle and be advanced at an angle of 60 degrees above the plane of the skin. If the nondominant hand is able to delineate the course of the common carotid artery, then the needle should to be advanced along a similar line just lateral to the carotid artery because both vessels are contained within the carotid sheath. If the carotid artery cannot be palpated with the nondominant hand, then the needle should enter the skin at the apex of the triangle at a 60-degree angle to the skin and be advanced in the direction of the ipsilateral nipple. If the needle is inserted 3 cm without achieving good blood return, then it should be pulled back slowly while applying constant backpressure to the plunger of the syringe and redirected more medially before being slowly advancing again. After the vessel has been cannulated, the modified Seldinger technique is used to complete insertion of the CVC (Fig. 12-2).

Posterior Internal Jugular Approach

The posterior internal jugular approach is an alternative to the central internal jugular approach that may be used if there is concern that the patient will not be able to tolerate a procedure-related pneumothorax (high positive end-expiratory pressure and/or high FiO2 requirements). The puncture site is posterolateral to the sternocleidomastoid muscle and immediately cephalad to where the sternocleidomastoid is crossed by the external jugular vein. The needle should be directed beneath the muscle and advanced in an anterior and inferior direction toward the sternal notch. If blood return is not obtained, the needle should be pulled back and redirected slightly more posterior until venous blood is obtained. Because unintentional carotid artery puncture is more likely with this approach, a “finder needle” should be used prior to cannulation with a large-bore needle.

Ultrasound-Guided Placement

Ultrasound guidance to assist in the placement of CVCs is a subject of controversy. Mansfield and colleagues29 reported that ultrasound guidance did not affect the rate of complications or failures in 821 patients randomized to standard insertion procedures with anatomic landmark guidance versus ultrasound guidance for subclavian vein catheterization. It is noteworthy that this study used ultrasonography to locate the subclavian vein but did not use real-time ultrasound guidance for the actual venipuncture. In contrast, several studies have compared the use of real-time ultrasound with the use anatomic landmarks during the insertion of both subclavian and internal jugular CVCs. These studies showed decreased failure rates, decreased complication rates, and an increased rate of successful catheter placement on the first attempt with the use of real-time ultrasound.30,31 A recent review of CVC complications reported a 6% to 10% incidence of mechanical complications with the insertion of subclavian and internal jugular CVCs.6 Given the frequent use of CVCs in the ICU and the risk of mechanical complications, it seems prudent to use real-time ultrasound during the insertion of CVCs if it is available, especially for patients with coagulation disturbances or unclear anatomic landmarks.

Catheter-related infections (e.g., bloodstream infection, catheter colonization, or an exit-site infection) are thought to arise via several different mechanisms: Skin flora from the insertion site can migrate down the external surface of the catheter, the catheter hub can become infected with repeated manipulation, or hematogenous seeding of the catheter tip can result from a distant source of bacteremia.22 CVC-related infections are the most common cause of nosocomial bacteremia in critically ill patients.14 The incidence of hospital-acquired CVC-associated bloodstream infections is collected by the CDC's National Nosocomial Infection Surveillance System (NNIS) and is expressed as the number of bloodstream infections (BSI) per 1000 CVC-days. From 1992 to 2001, the rate of CVC-related bloodstream infections in adult ICUs ranged from 2.9 to 5.9 per 1000 catheter-days.15 Diagnosis of a CVC-related BSI requires clinical symptoms of bacteremia (fever >38°C, chills, or hypotension) without another apparent source and isolation of an organism from a peripheral blood culture with either a semiquantitative or quantitative culture of a catheter segment that yields the same organism and antibiotic sensitivities as the organism cultured from blood. In the semiquantitative culture method, the catheter segment is rolled on a culture plate and considered positive if there are greater than 15 colony-forming units (CFUs) of an organism. In the quantitative method, the catheter is processed in broth and sonicated, followed by plating the broth on a culture plate. A positive culture requires growth of greater than 103 CFUs.16 CVC-related BSI should be distinguished from catheter colonization, which only requires a positive semiquantitative or quantitative culture from a catheter segment. In addition to BSI and catheter colonization, a CVC can develop an exit-site infection, defined as erythema, tenderness, induration, or purulence within 2 cm of the catheter exit site.16

The majority of pathogens causing CVC-related bloodstream infection are skin flora, which suggests migration of bacteria down the catheter as the mechanism of infection. A study of 297 pulmonary artery catheters found that 80% of infected catheters showed concordance with organisms cultured from the skin at the insertion site.5 According to NNIS data from 1992 to 1999, approximately 50% of hospital-acquired bloodstream infections were caused by staphylococcal species. The most common organisms isolated were coagulase-negative staphylococci (37%), Staphylococcus aureus (13%), enterococcus (13%), gram-negative rods (14%), and Candida spp. (8%).15 Given the frequency and cost associated with treating catheter-related infections, there has been a great deal of research into reducing the rate of these infections. Several interventions implemented at the time of catheter insertion have been shown to reduce the rate of catheter-associated infections.

Skin Preparation

The use of antiseptic skin preparations prior to sterile draping and percutaneous placement of CVCs is a routine part of the procedure. Although povidone-iodine is the most commonly used skin antiseptic agent in the United States, a recent meta-analysis reported a 50% reduction in catheter-related bloodstream infections with the use of chlorhexidine-based solutions rather than povidone-iodine (risk ratio 0.49, 95% confidence interval [CI] 0.28–0.88).19 This meta-analysis included several different types of chlorhexidine gluconate solutions for the insertion of central venous, peripheral venous, peripheral arterial, and pulmonary artery catheters. Subset analysis indicated that most of the benefit appeared to come from the chlorhexidine gluconate alcohol solutions rather than chlorhexidine gluconate aqueous solutions. Additionally, a subsequent meta-analysis determined that the use of chlorhexidine for central catheter site care resulted in a 0.23% decrease in the incidence of death and savings of $113 per catheter used.20 Finally, it should be noted that there are little data about whether antibiotic resistance emerges with the use of antiseptic solutions.

Maximal Sterile Barriers

Meticulous attention to sterile technique is of paramount importance during the placement of CVCs. The use of maximal sterile barriers likely decreases the incidence of inadvertent contamination of gloves, guidewires, and other equipment in the CVC kit. The technique of employing maximal sterile barriers, including a full body sterile drape, sterile gloves and long-sleeved gown, and nonsterile cap and mask, has been shown to reduce the incidence of CVC-related infections. Mermel and colleagues5 found that pulmonary artery catheters (PACs) placed in the ICU with the use of maximal sterile barrier precautions developed fewer infections (15.1% versus 24.6%, p <0.01) when compared with PACs placed in the operating room without maximal sterile barrier precautions. Raad and colleagues21 found that the time to occurrence of catheter-related bloodstream infection was lengthened in cancer patients who had CVCs and peripherally inserted CVCs inserted with maximal sterile barriers as compared with patients who had catheters inserted with sterile gloves and a small sterile field (p <0.05). Another study investigated the impact of a 1-day course taken by first-year resident physicians on infection control practices and hands-on instruction for several common procedures (arterial puncture and placement of arterial lines and CVCs). Subsequently, the documented use of full-size sterile drapes increased from 44% to 65%, and the rate of catheter-related infections decreased from 4.51 infections per 1000 patient-days before the first course to 2.92 infections per 1000 patient-days 18 months after the course. This decrease in catheter-related infections was associated with an estimated cost savings of $63,000.22

Antimicrobial Catheter Treatment

Another approach to reducing the incidence of catheter-related infections is the use of catheters treated with antimicrobial agents. Catheters coated with chlorhexidine and silver sulfadiazine and minocycline and rifampin are currently available for clinical use. Compared with conventional catheters, these antimicrobial-treated catheters are associated with significant reductions in catheter-related bacteremia.23,24 A study by Maki and colleagues25 comparing conventional triple-lumen polyurethane catheters with catheters coated with chlorhexidine and silver sulfadiazine reported both a reduction in catheter colonization and a nearly fivefold reduction in bloodstream infection. A recent randomized trial of minocycline and rifampin versus chlorhexidine and silver sulfadiazine catheters, however, found a significant decrease in the incidence of catheter-related BSI in the group of patients using minocycline and rifampin (0.3% versus 3.4%, p <0.002).26 There are several possible reasons for the differences between the two catheters. The minocycline and rifampin catheters had antibacterial substances both outside and inside the catheters, as opposed to the chlorhexidine catheters, which were only externally coated. Importantly, neither study reported hypersensitivity reactions to the catheters, and neither reported the occurrence of infections by organisms with resistance to the antimicrobial agents. A newer chlorhexidine and silver sulfadiazine catheter with antiseptic located on both the internal and external surfaces is now available, and studies comparing this new catheter with the minocycline and rifampin catheter are underway. While the exact role of the catheters remains the subject of some debate, a cost-benefit analysis suggests that antiseptic catheters are beneficial if an institution's rate of catheter-related bacteremia is 3 infections per 1000 catheter-days.25

Catheter Exchange over a Guidewire

In order to avoid mechanical complications of placing a new CVC, a strategy of inserting a sterile guidewire through an existing catheter, removing the catheter, and inserting a new sterile CVC over the guidewire is employed frequently. Since the existing catheter is not sterile, contamination of the new catheter is a concern with this technique. There have been several studies comparing scheduled catheter exchange over a guidewire and scheduled replacement of a CVC at a new site every 2 days, 3 days, 7 days, or as needed. In a meta-analysis, Cook and colleagues27 found trends toward higher catheter colonization (relative risk [RR] 1.26, 95% CI 0.87–1.84) and catheter-related bacteremia (RR 1.72, 95% CI 0.89–3.33) associated with exchange over a guidewire. Additionally, prophylactic catheter replacement was not found to reduce catheter colonization or catheter-related bacteremia as compared with replacement of the catheter on an as-needed basis. In fact, the CDC strongly recommends that CVCs should not be replaced or exchanged over a guidewire on a routine basis15 (category 1B recommendation: strongly recommended for implementation and supported by some experimental, clinical, or epidemiologic studies and a strong theoretical rationale). However, most authorities recommend wire exchange for suspicion of infection or mechanical catheter dysfunction. If infection is suspected, the catheter should be cultured on removal, and the new catheter should be removed if the culture of the old catheter is positive (>15 CFUs by the roll-plate method). Catheters with inflamed or purulent entry sites should be removed, and a new catheter should be inserted into a new site6,28 (Fig. 12-3).

Lefrant and colleagues32 reported their experience with subclavian vein catheterization over a 5-year period. Seven hundred and seven patients in a surgical critical care unit had subclavian vein catheterization attempted, with 562 successful procedures (79.5%). For the remaining 145 catheterizations, there were 67 failed procedures (overall failure rate of 9.5%). By multivariate analysis, more than one attempted venipuncture was the only independent risk factor for failed catheterization and immediate complications (e.g., arterial puncture, pneumothorax, misplacement of catheter). Elderly patients (>77 years of age) were more likely to have immediate complications but not failed catheterization. It is noteworthy that the operator's level of training and experience (junior, but not senior, residents were supervised by a critical care anesthesiologist) did not affect outcomes in the study, suggesting that adequately supervised physicians in training can perform CVC placement safely. Based on their observations, the authors recommended no more than two attempts at subclavian vein catheterization before aborting the procedure, with consideration of attempting at a different anatomic site. Contralateral attempts to cannulate the internal jugular or subclavian vein should be preceded by a chest radiograph to rule out pneumothorax, however, prior to proceeding.

Arterial Puncture/Bleeding

Accidental arterial puncture is a well-recognized complication of CVC placement. The incidence of this complication in published reports ranges from 0% to 15%.8–10 Complications arising from accidental arterial puncture include mediastinal hematoma formation, hemothorax, tracheal compression and possible asphyxiation, and retroperitoneal hemorrhage. A recent meta-analysis comparing internal jugular versus subclavian catheter placement noted a higher incidence of arterial puncture with internal jugular catheter attempts.10 Although arterial puncture occurs more frequently with internal jugular attempts, the carotid artery is more readily compressible compared with the subclavian artery, which makes this approach more attractive in patients with coagulation disturbances. Most complications of accidental arterial puncture occur with dilation and subsequent placement of large-bore catheters into an artery. Several case reports and small series have acknowledged this important complication.11–13 Traditional means of confirming arterial versus venous puncture of a blood vessel (e.g., bright red color, pulsatile blood return) may be unreliable in hypotensive, hypoxemic patients frequently encountered in the ICU. Transduction of the pressure waveform with intravenous extension tubing before dilation and placement of a large-bore catheter may reduce the occurrence of this complication. One may use tubing filled with sterile saline and attached to a three-way stopcock. After a vessel is entered, this tubing is connected to the needle while it is in the vessel, the tubing is elevated, and movement of the column of saline is analyzed to reflect either a venous or arterial waveform. Alternatively, the guidewire can be placed through the needle into the vessel using the modified Seldinger technique. Subsequently, a small, short catheter (e.g., 18- or 20-gauge 2-in intravenous catheter) can be placed over the wire into the vessel and the wire removed. The saline-filled, intravenous extension tubing can be attached to the catheter in the vessel to confirm a venous or arterial waveform. After confirmation of a venous waveform, the guidewire is replaced through this small intravenous catheter, the catheter is removed, and the procedure is finished. If unintentional arterial cannulation occurs, the small catheter is removed from the artery, and pressure is held at the site. Such a small catheter is much less likely to cause serious complications compared with a dilator and large-bore catheter. In our teaching hospital, we stress the importance of placing sterile intravenous extension tubing on the sterile field before the procedure is started so that a venous waveform can be confirmed prior to dilation and insertion of the catheter.

Bleeding complications from both arterial and venous punctures are exacerbated dramatically in patients who are thrombocytopenic or in those with coagulation disturbances. Unfortunately, such problems are common in critically ill patients. Those with platelet counts below 50,000/μL or those with an international normalized ratio (INR) above 2 probably should have catheters placed at a site with compressible vessels (e.g., internal jugular or femoral vein) unless the clotting problem can be corrected. The external jugular vein is an alternative that should be considered in those with clotting disturbances because it is a superficial vein that is easily compressible.

Pneumothorax

Pneumothorax is another important mechanical complication of CVC placement. The reported incidence of this complication ranges from 0% to 4.5%. Although some studies have reported a higher incidence of pneumothorax with subclavian catheter placement, a recent meta-analysis did not describe differences in the incidence of this complication when internal jugular and subclavian approaches were compared.10 Although many pneumothoraces occurring after CVC placement may not require treatment,33 patients undergoing positive-pressure ventilation should have the pneumothorax evacuated. The use of small-caliber pleural catheters is effective as an alternative to conventional tube thoracostomy in evacuating simple iatrogenic pneumothoraces34; however, since this therapy has not been tested extensively in patients undergoing positive-pressure ventilation, tube thoracostomy remains the treatment of choice in this situation.

Thrombotic Complications

Catheter-related venous thrombosis is a relatively common complication, occurring in between 2% to 66% of catheters.5,39 Catheter-related venous thrombosis may manifest as either a fibrin sleeve around the catheter or a thrombus that adheres to the wall of the vein that is typically asymptomatic. Because CVCs injure the endothelium and expose the venous intima, the coagulation system can become activated, resulting in thrombus formation. Difficulty with insertion of the line appears to increase the incidence of thrombosis, presumably owing to a greater degree of local venous trauma.40 There is some evidence that in an ICU setting a subclavian vein CVC is less likely to develop a catheter-related thrombosis than an internal jugular vein CVC.41 Timsit and colleagues41 used color Doppler ultrasound just before or within 24 hours of catheter removal to determine the frequency of catheter-related thrombosis associated with 208 CVCs placed in the ICU (catheters in place for 9.35 ± 5.4 days). A catheter-related internal jugular or subclavian vein thrombosis occurred in 42% (95% CI 34–49) and 10% (95% CI 3–18), respectively. The overall rate of thrombus formation was 33%, and an internal jugular CVC increased the risk of thrombus formation by a factor of four (RR 4.13 [95% CI 1.72–9.95]). Importantly, this study also determined that the risk of catheter-related sepsis was 2.62-fold higher when thrombosis occurred (p = 0.011). These findings contradict two studies that examined the incidence of complications in more permanent tunneled catheters and found a decreased incidence of venous stenosis and thrombus formation in the internal jugular group as compared with the subclavian group.42,43 Finally, the femoral vein is the least desirable anatomic location with regard to the risk of venous thrombosis. Merrer and colleagues4 reported that 25 of 116 (21.5%) patients randomized to femoral vein catheterization had ultrasound-detected venous thrombosis. This differed dramatically from those randomized to subclavian vein catheterization, in whom 2 of 107 (1.9%) had venous thrombosis (p <0.001).4

Catheter Occlusion

Occlusion of a CVC is another mechanical complication that occurs, particularly when catheters have been in place for extended periods of time. Thrombosis at the tip of the catheter may lead to this problem. Other reasons for CVC occlusion include precipitation of incompatible medications, a problem that can be avoided by careful attention to medication compatibility. Patients may suffer occasionally from the “pinch-off syndrome,” in which a subclavian catheter is compressed between the clavicle and the first rib.17 This complication usually occurs in long-term indwelling catheters; however, a narrow space between the clavicle and the first rib sometimes can interfere with successful placement of subclavian catheters, particularly those of large bore. CVCs placed for extended time periods have been reported to break and embolize to the right side of the heart or pulmonary artery, requiring radiologic or surgical removal.17,33

Catheter Misplacement

Most CVC placements in critically ill patients are performed without direct real-time visualization of the catheter. Typically, a chest radiograph is obtained after the procedure to ensure proper catheter position and to assess for evidence of a pneumothorax. Malpositioned catheters can then be repositioned correctly. Ideally, thoracic CVCs should terminate in the superior vena cava. Catheters occasionally may terminate in subclavian or jugular vein, as well as in the azygous, internal mammary, or pericardiophrenic vein, which may result in vascular injury and even perforation. When the tip of a catheter is positioned in the right atrium or right ventricle, perforation and subsequent cardiac tamponade may result.35,36 In order to avoid complications when CVCs are positioned in cardiac chambers, it is recommended that the tip of the catheter lie proximal to the angle between the trachea and the right mainstem bronchus.37 Ultrasonic examination after CVC placement may provide an alternative means of assessing adequacy of catheter placement.44

Air Embolism

When there is a communication between the great veins and the atmosphere, air may enter into the venous system. This potential complication is particularly relevant when considering the large-bore venous catheters used frequently in critically ill patients. Dysfunctional one-way valves or uncapped catheters may allow air to enter the venous system when intrathoracic pressure is subatmospheric during inspiration. Another problem is the possibility of venous air embolism during catheter removal, when a communication from the skin to a great vein may occur temporarily. Use of the Trendelenburg position and bioocclusive dressings may prevent this problem.38

In conclusion, intravascular catheters are necessary routinely for the management of critically ill patients. Mechanical, infectious, and thrombotic complications contribute considerable morbidity and mortality to these vulnerable patients. Recent evidence suggests that the subclavian vein may be the most desirable anatomic location for CVC placement; however, thrombocytopenia or coagulation disturbances—common problems in critically ill patients—may preclude this approach in some patients. It is encouraging that evidence to guide the appropriate management of CVCs is accumulating. Such evidence should allow clinicians to use these potentially lifesaving devices effectively while minimizing complications associated with their use.