a b s t r a c t

Purposes: We aimed to analyze the diagnostic accuracy of contrast enhanced ultrasonography (CEUS), compared with Chest x-ray , in the detection of correct Central venous line (CVL) placement. Our hypothesis was to verify whether CEUS could substitute CXR as a reference standard for correct placement of CVL or function as a triage test to limit the execution of CXR only for selected patients.

Basic Procedures: CEUS was carried out in 71 non consecutive patients to verify the correct positioning of a central venous line. Sensitivity, specificity, positive and negative predictive values, and likelihood ratios of CEUS compared to CXR, with their respective 95% confidence interval (CI), were calculated.

Main Findings: CXR identified 6 CVL misplacements (8,4%, CI 95% 3,2%-18%). Four of these were intravascular and 2 in the right atrium. CEUS identified only 3 misplacements, of which 1 was intravascular and 2 intracardiac. Using CXR as a reference standard, and considering intravascular and intracardiac malposition- ing altogether, the sensitivity, specificity, and positive and negative likelihood ratio of CEUS were 33% (95% CI, 0%-71%), 98% (95% CI, 95%-100%), 21%, 7%, 0%, and 68%, respectively. The negative and positive predictive

values were 94% (95% CI, 89%-100%) and 67% (95% CI, 13%-100%).

Principal conclusions: CEUS can’t substitute CXR, or become a triage test in selected patients, in evaluating the correct tip position after CVL placement.

(C) 2013

Introduction

central venous catheterization of subclavian (SC) and internal jugular veins (IJV) is associated with a 5% to 19% rate of early mechanical complications and misplacements [1,2]. Chest radiogra- phy (CXR) is usually carried out after central venous line (CVL) placement to exclude mechanical complications and to check the correct central venous line position. However, CXR is not always promptly available and this could cause dangerous delays in emergency. Moreover CXR necessitates a technician and has relevant biological costs in children and pregnant women. Several methods have been proposed to locate the tip of catheter [3-6], but at present CXR remains the exam used in daily practice and it is considered as the reference standard [7]. US guidance has become a standard of care for IJV placement [8-11], and it has been shown to be safer and to increase success rate of subclavian vein cannulation in comparison to the landmark method [12]. If, in addition to a high diagnostic accuracy in

* Corresponding author. Unita Operativa di Pronto Soccorso e Medicina D’urgenza, Dipartimento Di Medicina di Emergenza e Urgenza, Ospedale Maggiore Niguarda, Milano, Italy. Tel.: +39 3477938979.

E-mail address: [email protected] (F. Cortellaro).

recognizing mechanical complications [13-15], ultrasound would be proven to be accurate in identifying the correct position of CVL tip, it could be proposed as a substitute for CXR after CVL positioning. Few studies have investigated ultrasound accuracy in recognizing the correct placement of CVL, and there is still no agreement on the best US technique to identify the CVL tip position [16-18]. The aim of this study was to analyze the diagnostic accuracy of contrast enhanced ultrasonography (CEUS), compared with CXR, in the detection of correct CVL positioning. Our hypothesis was that CEUS could substitute CXR as the reference standard for correct placement of CVL or function as a triage test to limit the execution of CXR only to selected patients.

Methods

Patients with an indication to CVL placement were prospectively enrolled when the physicians who participated in the study were on duty, in an emergency department of a large Italian city hospital.

A written informed consent about the procedure and the processing of personal data was obtained for each patient. The indication for CVL placement was based on the physicians’ clinical judgment. Exclusion criteria were age b 18 years and pregnancy.

0735-6757/$ - see front matter (C) 2013 http://dx.doi.org/10.1016/j.ajem.2013.10.001

F. Cortellaro et al. / American Journal of Emergency Medicine 32 (2014) 78-81 79

Fig. 1. A, Subcostal acoustic window. B, Apical acoustic window. LA, left atrium; LV, left ventricule.

Three-way Blue flexTip catheters (Arrow-Howes Reading, PA) with 7F diameter and 20 cm length were used. CVL placement was carried out under US guidance with the standard Seldinger technique and the correct vessel positioning of the guidewire was checked in long axis before dilatation [7]. After insertion, all catheters were secured with surgical stitches. At the end of the procedure lung US and CEUS were carried out while waiting for the CXR. US examinations were performed by the same operator who placed the CVL. The operators were 2 emergency physicians, one anesthetist expert in CVL placement and emergency ultrasound and one resident. The latter had previously performed under supervision 10 CVL positioning and 150

Table 1

Interpretation of microbubbling test, modified from A. Vezzani [16]

Characteristics Interpretation

No bubbles Negative test: possible extravascular, extracardiac misplacement

Few bubbles or appearance N 2 seconds Negative test: intravascular

misplacement in neck veins or tip position too far from RA

Fig. 2. CEUS: visualization of micro-bubbles in the right atrium. Arrows indicate micro- bubbles coming from superior vena cava.

US examinations, according to American College of Emergency Physicians guidelines [9]. B-Mode cardiac ultrasound examinations were carried out using a 2- to 2.5-MHz sector probe (Esaote, My Lab 25-Medical System, Firenze, Italy) through either the SC or apical (A) acoustic windows when the SC window was not obtainable (Fig. 1A and B). After direct visualization of superior vena cava (SVC), right atrium (RA) and right ventricle (RV) at the same time, CEUS was performed in conformity with the standard technique used by cardiologists for detecting intra-cardiac shunt by preparing a 10 mL syringe containing 9 mL of saline solution and 1 mL of air, mixed with a 3-way stopcock to obtain a homogeneous solution [19]. We injected a 5 mL bolus through the distal lumen of the catheter and observed the pattern of micro-bubbles flow entering the RA. CEUS was performed always from the distal port of the CVL: after insertion we checked the correct functioning of the three lines, sucking blood in a siringe from each line, after this procedure we injected contrast through the distal port. Recorded clips of the CEUS examinations within 3 heart-beats were reviewed and, in case of doubt, a second Bolus injection of 5 mL was repeated. Interpretation of CEUS was made according to the criteria described in Table 1 [16]. SVC and right atrium visualization (study feasibility), time to carry out the procedure and concordance with CXR were recorded and analyzed. Inter-observer variability of CEUS results among the three operators was assessed by blindly reviewing 20 clip images. All operators were blinded to CXR results. CXR was acquired in a supine position (Siemens Rohere/Tube Optitop 150/40/80 HC 2007) and was examined by the radiologist on duty who was blind to the CEUS result. CVL misplacement was defined as evidence of the tip in the right atrium or right ventricle (intra-cardiac misplacement), in the controlateral neck veins or in the homolateral opposite veins (SC if positioned from IJV or vice versa) or in the brachiocephalic veins

Table 2

Characteristics of the patients and CVL placement

71 patients

Female sex, N (%) 27 (38)

Numerous bubbles with turbulent flow in the RA or direct visualization of catheter tip in right atrium

Numerous bubbles with linear flow coming from SVC within 2 seconds

Negative test: intracardiac (RA) misplacement

Positive test: correct tip positioning

Age (mean +- SD) 71 +- 10

Right IJV approach, N (%) 44 (62)

Left IJV approach, N (%) 17 (24)

Right subclavian approach, N (%) 7 [10]

Left subclavian approach, N (%) 3 [4]

80	F. Cortellaro et al. / American Journal of Emergency Medicine 32 (2014) 78-81
Table 3A Concordance between CEUS and CXR
	CXR		Total
	Incorrect placement	Correct
	intra-atrial intravascular	placement
CEUS Incorrect placement	intra-atrial 1 0	1	2
	intravascular 0 1	0	1
Correct placement	1 3	64	68
Total	2 4	65	71

Table 3B

Concordance between CEUS and CXR for intravascular misplacement

Chest x-ray Total

Incorrect placement Correct placement

CEUS Incorrect placement 1 0 1

Correct placement 3 67 70

Total 4 67 71

For sensitivity and specificity see text.

(intravascular misplacement). A true incorrect placement was defined as the absence or the appearance of only a few bubbles or late (N 2 seconds) echo-contrast entering the RA from the SVC or turbulent flow in the RA or RV, confirmed by CXR (Fig. 2). A false incorrect placement was defined as the absence or the appearance of only a few bubbles or late (N 2 seconds) echo-contrast entering the RA from the SVC or turbulent flow in the RA or RV with CXR evidence of catheter in the SVC. A false correct placement was defined as the presence of laminar flow into the RA and CXR evidence of the catheter in the neck veins (IGV, brachiocefalic, SC) or right heart. A true correct placement was defined as the presence of laminar flow from SVC into the RA, confirmed by CXR. Sensitivity, specificity, positive and negative predictive values and likelihood ratio of CEUS compared to CXR, with their respective 95% confidence interval, were calculated. Since the required sample size to reach a lower boundary of the confidence interval for sensitivity higher than 90% would have required a huge number of subjects we decided to run a pilot study to see if our hypothesis could deserve an ad hoc trial. Global concordance and ? statistics were calculated for the inter-observer variability of CEUS results. All analyses were carried out using Med Calc (MedCalc Software 12.3, Broekstraat 52, 9030 Mariakerke, Belgium). The present study was approved by the Ethical Committee of our hospital.

Results

Seventy one patients were enrolled. The main characteristics of the population are shown in Table 2. CEUS was feasible in all patients and a good acoustic window was obtained in all patients either with a subcostal (59 patients) or apical (12 patients) view, in all cases the SVC-RA junction was clearly visible allowing to identify pattern of CEUS in all patients within 5 minutes (mean time 4 +- 1 min). Mean

time to CXR results (CXR read by the radiologist on duty) was 288 +- 216 min. Results of intravascular or intra-cardiac tip detection with CXR or CEUS are shown in Table 3. CXR identified 6 CVL mis- placements (8.4%, 95% CI, 3.2%-18%). Four of these were intravascular (two in the brachiocephalic vein, one in the SC vein, one in the IJV) and two in the RA. CEUS identified only three CVL misplacements (4.2% 95% CI, 1.5%-11.7%) of which one was intravascular and two intracardiac in the RA (video 1 shows RA malpositioning). In one case the incorrect intravascular placement was identified by both techniques. CEUS showed absence of flow in the RA and CXR identified CVC in IJV coming from the SC vein. There were three cases of false correct intravascular placement (see Table 4), in one case of insertion in the left IJV CXR showed the tip in the right brachiocephalic vein (CXR: Fig. 3, CEUS: video 2). Regarding the 2 intracardiac malpositions, there was discordance in the tip position only in one case (CEUS considered the tip to be in the RA while it resulted to be in the SVC after CXR). Using CXR as a reference standard, and considering intravascular and intracardiac malpositions altogether, the sensitivity, specificity, positive and negative likelihood ratio of CEUS were 33% (95% CI, 0%-71%), 98% (95% CI, 95%-100%),

21%,7%, 0%, and 68%, respectively. The negative and positive predictive values were 94% (95% CI, 89%-100%), 67% (95% CI, 13%-100%)

(Table 3A). For intravascular misplacement detection, sensitivity and specificity for CEUS were 25% (95% CI, 1%-78%) and 100% (95% CI, 93%-100%), respectively (Table 3B).

Interobserver concordance between recorded clips of the CEUS examination was calculated on 25 patients. Global concordance was 96% (95% CI, 82%-100%) with a ? statistic of 0.65 (95% CI, 0.01-1). There was discordance in one case of SCV malposition from IJV insertion: CEUS was judged normal by the first operator and delayed- doubtful by the second operator (video 2). Errors were uniformly distributed between all the operators.

Discussion

In the present study we aimed to verify whether CEUS could substitute CXR as the reference standard for CVL placement or could serve as a triage test to limit CXR only to doubtful cases. The results of our study lead to the opposite conclusion: CEUS by itself cannot substitute CXR or serve as a triage test. Indeed, its sensitivity is 33% and, also considering the best scenario of the upper CI limit, it would still be lower than 71%. Other studies have investigated a possible role

Table 4

Characteristics of patients not recognized by CEUS

Patients	Clinical presentation	BMI	Line access site	CEUS results	CEUS time of visualization	interpretation	CXR results
1	Hematemesis in cirrhosis	22	Right IJV	laminar flow	b2 s	Correct tip placement	Right subclavian vein
2	Hemoperitoneum in cirrhosis	31	Left IJV	laminar flow	b2 s	Correct tip placement	Right brachiocephalic vein
3	Urosepsis	23	Left IJV	laminar flow	b2 s	Correct tip placement	Left
							brachiocephalic vein
4	Pneumonia	24	Right IJV	laminar flow	b2 s	Correct tip placement	Right atrium

Shows the characteristics of the malpositions not recognized by CEUS. In particular only one subject had a body mass index more than 25. BMI, body mass index.

F. Cortellaro et al. / American Journal of Emergency Medicine 32 (2014) 78-81 81

Limitations

Fig. 3. Chest XR: visualization of central venous line’s tip in the right brachiocephalic vein (arrow).

for ultrasonography in place of CXR for CVL position control. Zanobetti et al. tried to visualize directly the catheter at the end of the procedure in the neck vessels [18]. They found a good correlation between CXR and US (? = 0.82); however, they assumed that the catheter tip was at the level of the superior vena cava even in the absence of direct visualization. Indeed, ultrasound has the intrinsic limitation that it cannot detect the correct placement of the catheter tip inside the thoracic cavity (from the clavicula to the SVC-RA junction). This means that the tip might be anywhere distal the brachiocephalic vein. In our study this implies that 3 out of the 4 cases of intravascular malpositioning would have not been recognized with ultrasound direct visualization of the tip. Lanza et al. [17] studied 107 children in the intensive care unit comparing CXR with B-mode and Color Doppler US and found the sensitivity of US to be 85% and its specificity 100%. Their final suggestion was that CXR should be limited only to selected cases, when inappropriate catheter tip position or pneumothorax are suspected after sonography. It must be noted that they studied a population of children in whom the risk of CXR in terms of radiation exposure is higher. Since CVL misplacement can have serious complications, we think that a good substitute for CXR should have at least a similar sensitivity in its identification. We wonder whether 85% of sensitivity is good enough and whether their results would be reproducible in a population of adults. Finally, Vezzani et al. [16] compared CXR with CEUS and B-mode in 111 adult patients. They found a 96% sensitivity and 93% specificity of US for CVL misplace- ment, but intravascular malpositionings were all in the IJV, and were identified only by B-mode US and not by CEUS. It is not known whether they would have been able to recognize malpositionings in the SVC or brachiocephalic veins. Even if our study does not support the use of CEUS to control the position of the catheter, US has nevertheless some advantages compared to CXR: it can be done quickly, there is no need for a technician, it has no biological costs, it can be done by the same operator who places the CVL, it can be easily learned and it is more accurate than CXR in identifying mechanical complications of CVL placement [15]. Even if CEUS alone does not have an adequate sensitivity, we think that more studies should be done, combining CEUS and the direct visualization of the guidewire during cannulation, or of the catheter after cannulation in the neck veins. We think that these studies should clearly report US sensitivity compared to CXR especially for intravascular malpositioning and should clearly report the feasibility of US. Since the total number of misplacements in our and other studies is low, the negative predictive value of CEUS is relatively high. Thus in emergency CEUS could still be a useful adjunct while awaiting to carry out CXR.

Some limitations of our study should be addressed. First of all, we enrolled a limited number of patients; however, the upper limit of the sensitivity confidence Interval is too low to be clinically useful. Furthermore, we enrolled only patients who presented when the physicians involved in the study were on duty. This could constitute an enrollment bias. However, the patients were enrolled consecu- tively and there is no reason to think that patients who presented at different times would be different in any respect. Finally, as CEUS could be more accurate in the pediatric population, a potential benefit to this group should not be excluded by inference of the current results. In conclusion, our study shows that CEUS cannot substitute CXR in evaluating the correct catheter tip position after CVL placement. Moreover, CEUS cannot be a triage test to limit CXR only to selected cases. More studies are needed to evaluate whether combining CEUS with other US techniques could increase diagnostic accuracy. Till then CEUS could only be used to confirm intravenous positioning and start an infusion when medications must be given without delay, while waiting for CXR.

Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ajem.2013.10.001.

References

1. McGee DC, Gould MK. Preventing complications of central venous catheterization. N Engl J Med 2003;348:1123-33.
2. Gladwin MT, Slonim A, Landucci DL, et al. Cannulation of the internal jugular vein: is postprocedural chest radiography always necessary? Crit Care Med 1999;27: 1819-23.
3. Andropoulos DB, Stayer SA, Bent ST, et al. A controlled study of transesophageal echocardiography to guide Central venous catheter placement in congenital heart surgery patients. Anesth Analg 1999;89:65-70.
4. Parigi GB, Verga G. Accurate placement of central venous catheters in pediatric patients using endocavitary electrocardiography: reassessment of a personal technique. J Pediatr Surg 1997;32:1226-8.
5. Ender J, Erdoes G, Krohmer E, et al. Transesophageal echocardiography for verification of the position of the electrocardiographically-placed central venous catheter. J Cardiothorac Vasc Anesth 2009;23:457-61.
6. Rath GP, Bithal PK, Toshniwal GR, et al. Saline flush test for bedside detection of misplaced subclavian vein catheter into ipsilateral internal jugular vein. Br J Anaesth 2009;102:499-502.
7. American Society of Anesthesiologists Task Force on central venous access, Rupp SM, Apfelbaum JL, Blitt C, et al. Practice Guidelines for Central Venous Access, a Report by the American Society of Anesthesiologists Task Force on Central Venous Access. Anesthesiology 2012;116:539-73.
8. Neri L, Storti E, Lichtenstein D. Toward an ultrasound curriculum for critical care medicine. Crit Care Med 2007;35:S290-304.
9. American College of Emergency Physicians. emergency ultrasound guidelines. Ann Emerg Med 2009;53:550-70.
10. Lamperti M, Bodenham AR, Pittiruti M, et al. International evidence-based recommendations on ultrasound-guided vascular access. Intensive Care Med 2012 Jul;38:1105-17.
11. Troianos CA, Hartman GS, Glas KE, et al. Councils on Intraoperative Echocardiog- raphy and Vascular Ultrasound of the American Society of Echocardiography. Guidelines for Performing ultrasound guided Vascular Cannulation: Recommen- dations of the American Society of Echocardiography and the Society of Cardiovascular Anesthesiologists. J Am Soc Echocardiogr 2011;24:1291-318.
12. Fragou M, Gravvanis A, Dimitriou V, et al. Real-time ultrasound-guided subclavian vein cannulation versus the landmark method in critical care patients: a prospective randomized study. Crit Care Med 2011;39:1607-12.
13. Tocino IM, Miller MH, Fairfax WR. Distribution of pneumothorax in the supine and semirecumbent critically ill adult. Am J Roentgenol 1985;144:901-5.
14. Jalli R, Sefidbakht S, Jafari SH. Value of ultrasound in diagnosis of pneumothorax: a prospective study. Emerg Radiol 2013;20:131-4.
15. Volpicelli G, Elbarbary M, Blaivas M, et al. International Liaison Committee on Lung Ultrasound (ILC-LUS) for International Consensus Conference on Lung Ultrasound (ICC-LUS). Intensive Care Med 2012 Apr;38(4):577-91.
16. Vezzani A, Brusasco C, Palermo S, et al. Ultrasound localization of central vein catheter and detection of postprocedural pneumothorax: an alternative to chest radiography. Crit Care Med 2010;38:533-8.
17. Lanza C, Russo M, Fabrizzi G. central venous cannulation: are routine chest radiographs necessary after B-mode and colour Doppler sonography check? Pediatr Radiol 2006;36:1252-6.
18. Zanobetti M, Coppa A, Bulletti F, et al. Verification of correct central venous catheter placement in the emergency department: comparison between ultraso- nography and chest radiography. Intern Emerg Med 2013;8:173-80.
19. Jauss M, Zanette E. Detection of right-to-left shunt with ultrasound contrast agent and transcranial Doppler sonography. Cerebrovasc Dis 2000;10:490-6.