Verification of correct central venous catheter placement in the emergency department: comparison between ultrasonography and chest radiography
In 210 consecutive patients undergoing emergency central venous catheterization, we studied whether an ultrasonography examination performed at the bedside by an emergency physician can be an alternative method to chest X-ray study to verify the correct central venous catheter placement, and to identify mechanical complications. A prospective, blinded, observational study was performed, from January 2009 to December 2011, in the emergency department of a university-affiliated teaching hospital. Ultrasonography interpretation was completed during image acquisition; ultrasound scan was performed in 5 ± 3 min, whereas the time interval between chest radiograph request and its final interpretation was 65 ± 74 min p < 0.0001. We found a high concordance between the two diagnostic modalities in the identification of catheter position (Kappa = 82 %, p < 0.0001), and their ability to identify a possible wrong position showed a high correlation (Pearson’s r = 0.76 %, p < 0.0001) with a sensitivity of 94 %, a specificity of 89 % for ultrasonography. Regarding the mechanical complications, three iatrogenic pneumothoraces occurred, all were correctly identified by ultrasonography and confirmed by chest radiography (sensitivity 100 %). Our study showed a high correlation between these two modalities to identify possible malpositioning of a catheter resulting from cannulation of central veins, and its complications. The less time required to perform ultrasonography allows earlier use of the catheter for the administration of acute therapies that can be life-saving for the critically ill patients.
KeywordsEmergency medicine Emergency ultrasonography Chest radiography Central venous catheter
The central venous catheterization, performed by cannulation of subclavian or internal jugular veins, is currently a common procedure in the care of critically ill patients, both in the emergency department (ED) and in the Intensive Care Unit (ICU). This procedure represents an essential means for the monitoring of central venous pressure, the administration of drugs, fluids, blood products and total parenteral nutrition. Despite its widespread use, the placement of a central venous catheter (CVC) remains a procedure followed by complications in a rate ranging from 0.3 to 12 % depending on operator’s experience and definition of a complication itself . Even when the procedure is performed by an expert operator, a CVC can advance in the wrong vessel. The main mechanical complications relate to vessel puncture and catheter advancement are pneumothorax, vascular complications such as arterial puncture and hematoma at the insertion point, and damage of neural structures . Unlike these early mechanical complications, at present, serious complications like infections or thrombosis are relatively uncommon [1, 3, 4, 5]. The correct position of the catheter tip remains a controversial issue. Its localization in the superior vena cava, just above its outlet into the right atrium is usually considered to be the ideal position for minimizing the risk of cardiac tamponade and venous thrombosis . After the CVC placement, a standard chest radiography (CXR) is recommended to verify the CVC position, and to rule out possible complications. Although several other methods have been proposed to locate the tip of the catheter and to exclude any malposition, such as endocavitary electrocardiography, transesophageal echocardiogram, or flush with saline testing [7, 8, 9, 10], the CXR is currently considered the gold standard. However, its accuracy in locating the catheter tip is certainly overestimated . Despite the simplicity of this diagnostic modality, the entire procedure of a CXR, from the examination request to the final report, requires a considerable time interval that can be very harmful to the critically ill patient. The use of ultrasound (US) guidance, which has become a common practice in particular for transjugular and peripheral CVC insertion, has made this procedure safer and increased the success rate . As recently demonstrated, a focused ultrasonography aims to provide emergency response to specific questions, and performed at the bedside by an emergency physician (EP) or an intensivist, can be an extremely useful tool in caring for a critically ill patient [13, 14]. The US examination of blood vessels, heart and chest performed by a non-radiologist physician is a feasible diagnostic modality with many clinical advantages [15, 16, 17, 18]. However, despite its growing use, there are only few scientific papers demonstrating that transthoracic US is as accurate as CXR in defining the position of the CVC tip [19, 20, 21, 22], and all these studies were carried out in the ICU. While the transthoracic US accuracy in detecting the presence of pneumothorax [16, 23] as well as the ease in the display of the inferior vena cava, the subclavian and the internal jugular vein has been demonstrated [24, 25], the visualization of the superior vena cava, particularly in its lower portion at the level of its junction with the right atrium, is sometimes difficult. The aim of our study was to determine whether, in a series of patients undergoing emergency central venous catheterization, the US performed at the bedside by an EP can be an alternative method to CXR to verify the correct placement of the catheter and to diagnose mechanical complications such as pneumothorax, pleural and pericardial effusion.
A prospective, blinded, observational study was performed. The study consistent with the principles of Declaration of Helsinki on clinical research involving human subjects was approved by an ad hoc ethics committee. For each patient, written informed consent about the procedure and the processing of personal data were obtained. Patients were enrolled, from January 2009 to December 2011, in the ED of an urban academic level I trauma center and a tertiary-care facility with an annual ED census of 70,000 visits. The ED is the primary teaching site for the emergency medicine residency. The indication for central venous catheterization was based on clinical judgment by EPs participating in the study. The most common indications included central venous pressure monitoring, rapid restoration of blood volume and administration of drugs requiring central venous delivery. Patients younger than 18 years were excluded from the study.
Three-way Blue FlexTip catheters (Arrow-Howes, Reading, PA, USA) with 7 French diameter and 20-cm length were used. The catheter placement was performed by a senior EP or by an emergency medicine resident. The procedures performed by in-training physicians were carried out under the tutor supervision. The approach, transjugular or subclavicular, was chosen by the operator according to the clinical situation, the characteristics of the patient and personal preferences. The catheters were inserted with the patient in a supine position under sterile conditions using the Seldinger technique. The subclavian vein use a subclavicular approach based on the recognition of anatomical landmarks; access in the internal jugular vein was instead performed by the use of the ultrasonographic guide. Immediately after CVC insertion, an US B-scan was performed to identify the catheter tip position and to rule out any mechanical complications. The US examination was performed by the same operator who had placed the CVC. To acquire the necessary skill to perform the US examination, EPs and residents attended a theoretical training on vascular, cardiac and chest ultrasonography lasting 8 h for each area, and performed, under a tutor supervision, 25 ultrasonographic examinations of each kind, as suggested by American College of EP’ guidelines .
US examinations were performed with an iE33 ultrasound system (Philips Ultrasound, Bothell, WA, USA) equipped with a 4–8 MHz linear probe and a 2.5–3.5 MHz sector probe. The images were recorded, and in case of discordance with the CXR were revised. The US examination was performed with the patient in a supine position; if not contraindicated by the clinical condition, images from the apical window were acquired with the patients in a left lateral decubitus position. Internal jugular and subclavian veins were extensively scanned bilaterally by the linear probe from latero-cervical, supraclavicular and infraclavicular fossae to identify the CVC. Heart chambers and pericardial space were visualized from apical and subcostal windows by the sector probe to identify the presence of the catheter inside the heart, and to rule out pericardial effusion and tamponade. If the CVC was inserted through the internal jugular vein and the US scan was unable to visualize a tip entrance into the contra lateral jugular vein, into the subclavian veins bilaterally, and into the heart chambers, we assumed that the catheter tip was at the level of the superior vena cava. When the subclavicular access was used and the catheter was not visualized in the contralateral subclavian vein and at the level of the internal jugular veins bilaterally, and within the heart chambers, we similarly assumed that the catheter tip was in the superior vena cava. We considered as correct placements all cases in which the catheter tip position has been defined at the level of the superior vena cava or into the right atrium nearby the superior venae cava outlet; the finding of the catheter tip at the level of venous vessels other than the superior vena cava or too far distally into the right atrium has been considered as wrong positioning. Chest examination was performed with linear probe to check for the presence of complications such as pneumothorax and pleural effusion. Longitudinal and transversal scanning of both hemithoraces were performed along the parasternal, midclavicular, anterior, middle, and posterior axillary lines. The presence of pneumothorax was defined by the following ultrasound findings: absence of gliding sign, absence of vertical artifacts (B line); appearance of lung point . The presence of pleural effusion was identified as an hypoechoic area between the parietal pleura and visceral pleura.
At the end of the US scan, the physician performing the procedure filled out a card recording the following information: patient name and date of birth, admission diagnosis, drugs used for the procedural sedation, presence of chest alteration, insertion point; time required to perform the US scan; catheter tip position as identified by US exam, the presence of pneumothorax, hemothorax, hemopericardium with or without tamponade. The admission diagnoses were grouped into nine main categories: septic shock, trauma, lung disease, heart disease, abdominal disease, acute renal failure or metabolic or fluid disturbances, inability to obtain peripheral access, hypovolemic shock, or cardiopulmonary arrest. Chest alteration was defined as anatomic changes of the rib cage due to chronic obstructive pulmonary disease such as barrel chest, deterioration due to restrictive lung disease, chest trauma with rib fractures, scars of previous thoracic surgery (resection or lung transplantation) and alterations such as chest scoliosis or pectus carinatum. At the end of the CVC insertion, a bedside CXR was obtained to verify the catheter tip position, with the radiologist blinded regarding the US scan findings. The CXR was acquired with the patient in a supine position by a Philips 300 Mobile Practix radiographic system (Philips Healthcare, Amsterdam, The Netherlands) and developed with a Cr 85-X digitizer (AGFA Healthcare, Morstel, Belgium). The US findings were then compared with those obtained from the CXR, considered as the gold standard. Because the catheter position and the presence of complications were evaluated during the US scanning, the time required to get the US final report was considered equal to the time spent for the US examination. For the radiological examination, we calculated the time interval between the CXR request and the final report availability. US sensitivity, specificity, negative predictive value (NPV) and positive predictive value (PPV) were calculated with the CXR as gold standard. Continuous variables are reported as mean ± standard deviation and comparisons between the two modalities were performed using the t Student test for paired-data. To compare categorical variables, Fisher’s exact test was used. The concordance between ultrasonography and radiography was analyzed using the Cohen’s K test. A p value <0.05 was considered statistically significant.
All statistical analyses were performed using SPSS statistical package version 19.0 (SPSS Inc., Chicago, IL, USA).
Reasons for hospitalization
99 (47 %)
19 (9 %)
26 (12 %)
17 (8 %)
10 (5 %)
Acute renal failure and/or hydro-electrolytic disturbances
23 (11 %)
Inability to obtain peripheral access
5 (2 %)
9 (4 %)
2 (1 %)
Comparison of the position of catheter identified by ultrasonography with that showed on chest radiograph
Superior vena cava
Internal jugular vein
Superior vena cava
Internal jugular vein
Overall, ultrasonography was not able to correctly define the catheter position in relation to the right atrium in six patients. In one of the cases in which at the ultrasound examination the catheter was mistakenly placed in the right atrium, the patient had a dual-chamber pacemaker, and a previously inserted tunneled central venous catheter for dialysis. The remaining five cases had chest wall abnormalities (one patient with chest trauma and previously subjected to lung bilateral transplantation, two patients with barrel chest for chronic obstructive pulmonary disease, one patient with pectus excavatum, and one with chest alteration due to restrictive lung disease). We have subsequently evaluated the ability of ultrasound in identifying the catheter within the atrium in two subgroups of patients, those who had a normal conformation of the rib cage and those with anatomy alterations. In the first subgroup we have found a sensitivity of 98 %, a specificity of 93 %, a PPV of 93 % and a NPV of 98 %. The correlation between two methods was high (Kappa = 87 %, p < 0.0001). In the group of patients who had abnormal chest structure, sensitivity of US was found to be 75 %, specificity 64 %, PPV of 75 % and NPV of 64 %. In this subgroup, the correlation between the two techniques was only moderate (Kappa = 45 %, p = NS). Finally, to check whether the differences in the two groups were statistically significant and, therefore, attributable to the presence or absence of alterations of the structure of the rib cage, we compared the cases where the US was unable to correctly locate the catheter in relation to the presence of altered thoracic anatomy. The difference was statistically significant (p = < 0.041), demonstrating that the ability to identify if the catheter was located within the right atrium was affected by the presence of alterations of the anatomy of the chest.
The central venous catheterization is a safe procedure in most cases; but despite this, the wrong positioning of the catheter may be associated with even severe complications, including death of the patient. Some of these complications, such as the puncture bleeding, the development of hematoma or cervical nerve lesions, are easy and objective, while for the verification of the wrong positioning itself and the search for other complications such as pneumothorax, hemothorax or the hemopericardium, it is necessary to perform a chest X-ray as recommended by all protocols [28, 29]. The use of the catheter should be delayed until such time as it is certain that this is positioned correctly; often, however, the delay due to the time that elapses between the request of chest X-ray until reporting its results can be harmful to the critically ill patient. The time required to obtain the results of the X-ray is frequently very high, and in our series, the value was an average of 65 min ± DS 74. The time required to perform the US was found to be significantly shorter, with an average of 5 min ± DS 3, since with ultrasound diagnostic evaluation takes place during the acquisition of images itself. The time advantage of US is to be of great importance, particularly in emergency venous catheterization, for the opportunity to start treatment as early as possible greatly impacting on patient outcome, while in the case of the pending X-ray report, the patient cannot be handled properly, especially when mechanical complications have occurred. Some researchers have even questioned the real usefulness of the chest X-ray study performed after central venous catheterization in the absence of obvious clinical complications, whereas a simple procedure deemed by operator and a number less than three punctures to locate the blood vessel, has an high NPV for the presence of pneumothorax [30, 31], other studies have also highlighted the costs of this procedure [28, 32, 33].
A further topic against the execution of the chest X-ray study is represented by the exposure of the patient to ionizing radiation, especially for those admitted to the ICU where the daily recourse to the radiological diagnosis is high. The major technical advances of US that occurred in recent decades have led to considerable improvement in resolution of the images obtained with this method, so that the US is now a technique increasingly used in the field of emergency medicine and in ICU [34, 35, 36]. In fact, both the ability to perform the examination directly at the bedside, its non-invasiveness and repeatability are all features that make this method very advantageous in the management of critically ill patients, representing an important aid in the diagnosis, evaluation of the undertaken therapies and during invasive procedures. It has been demonstrated that the US has a specificity and sensitivity greater than the chest X-ray in the diagnosis of pneumothorax [16, 23, 34], especially in cases where the patient must maintain a supine obliged position. Similar data have also been obtained with regard to the diagnosis of hemothorax [34, 37]. With this method, we can also easily display both the jugular vein and the subclavian vein. Furthermore, the high echogenicity of the materials that make up the modern intravenous catheters makes their ultrasonographic identification particularly easy at the level of the main veins of the upper part of the body and also within the cardiac chambers [3, 34]. Although US is now widely used as a guide for insertion of central venous catheters, and this has led to an increased percentage of successful procedures [1, 24, 38] and to a reduction of time required for the procedure itself, so far, there are only few studies in the literature suggesting that this method for post-procedure evaluation is a possible alternative to chest X-ray [19, 20, 21, 22]. All these were made in ICU, in patients with different clinical characteristics and often undergoing mechanical ventilation; on the contrary, our population was formed of acutely ill patients who, in emergency situation, required the sudden insertion of a central venous line. Various studies have shown that US, limited to searching for the specific questions (focus directed), can be done by an EP without difficulty [34, 35, 36]. In all previous studies, it has always been necessary to have two operators for each procedure performed, one to insert the catheter and another to perform US. We thought that in clinical practice, it would be more advantageous, regarding the management of time and human resource, if both phases were carried out by one person. In our study, US was precisely performed by the EP who had placed the CVC, and was executable and interpretable in all cases contrary to the percentages of the feasibility of this method in the literature [20, 22]. Regarding the correct position of the CVC, the question appears to be still controversial. We believe that in case of short-term central venous catheterization, during medical emergencies, the correct position of the CVC is both at the level of terminal part of the superior vena cava, both inside the right atrium near the atrium–caval junction. In fact, in the literature, there are various studies which show that this position of the catheter tip is not associated with an increased risk of complications . Indeed, it is important to note that most of the literature case reports of cardiac perforation associated with intra-atrial catheter placement date back to the 1970s and 1980s, a time when the catheter materials were very different from those used today, resulting in a more rigid catheter with high thrombogenic intrinsic properties [40, 41]. Even the placement of the catheter at the level of the atrium associated with clinically significant complications such as venous thrombosis, it turns out to be much more frequent at this level when the catheter is located inside atrium [39, 42, 43]. The position at the level of the upper part of right atrium is, therefore, now considered safe by most researchers, provided that the tip is not in contact with the heart wall, and does not pass through the tricuspid valve [2, 39, 44, 45, 46]. Moreover, the venous oxygen saturation measured at the level of the atrium would seem to correlate more adequately to the mixed venous saturation measured into the pulmonary artery than that measured at the level of the superior vena cava . In our study, with US, we were able to identify five cases (2 %) of bad positioning of the catheter, which required a repositioning, showing a perfect correlation with the chest X-ray study. The catheter in all these cases was positioned at the level of the internal jugular vein ipsilateral to the cannulated subclavian. Regarding the mechanical complications, three cases of iatrogenic pneumothorax were detected, both shown by US and confirmed by chest X-ray study (sensitivity 100 %). The sensitivity and specificity of the US in identifying the intra-atrial position of the catheter were, respectively, 94 and 89 %, demonstrating a significant correlation with chest radiography, and these values were found to be higher in the subgroup of patients with a normal conformation of the thoracic cage (98 % sensitivity and specificity of the 93 %). The difference noted between the patients with abnormalities of the rib cage and those who showed normal anatomy appears to be statistically significant.
Among the limitations of our study is the fact that we compared US results with those obtained with chest X-ray, a technique which, though it is still considered the gold standard, in fact has a low accuracy in locating the catheter tip, and certainly much overestimated accuracy of the chest X-ray study. Indeed, the junction between the superior vena cava and right atrium is not directly visualized on chest X-ray study performed at the bedside. It has been demonstrated that the reading of the X-ray based on common radiological landmarks would lead from 20 to 40 % false-positive results as to the intra-atrial position of the tip of the catheter, and that none of the radiological landmarks used allows one to locate the catheter tip in a reliable manner to 100 % [48, 49]. In addition, inter-operator variability in reading the chest X-ray study is high . We must also remember that ultrasound does not identify the possible, though very rare, malpositioning of the catheter at the level of small veins (e.g., within the azygos vein), on the other hand, even a chest X-ray study is often not be able to demonstrate this complication . Our study showed that the US, performed at the bedside of patient in emergency conditions, is able to identify with accuracy and speed, possible malpositioning of the catheter resulting from cannulation of central veins, and its complications. A high correlation was found by comparison with chest X-ray study, still considered the gold standard, and the less time required to perform the US, justify the use of ultrasound as an alternative method to maintain control post-procedure, thus allowing early use of the catheter for the administration of acute therapies that can be life-saving for the critically ill patient. Moreover, the possibility that the same operator who has carried out the catheterization can check the positioning of the catheter and the possible presence of mechanical complications is very advantageous as regards the time management and human resources, resulting in a reduction of costs.
Conflict of interest
- 26.American College of Emergency Physicians (2009) Emergency ultrasound guidelines. Ann Emerg Med 53:550–570Google Scholar
- 44.National Kidney Foundation-Kidney Disease Outcomes Quality Initiative (2001) III. NKF-K/DOQI clinical practice guidelines for vascular access: update 2000. Am J Kidney Dis 37:S137–S181Google Scholar
- 45.Chamorro C, Pardo C, Alberto SJ, Borrallo Angel Romera Jose Manuel, Luis Martinez-Melgar J (2001) Malposition of central venous catheters in hospitalized patients. Med Clin (Barc) 117:12–13Google Scholar
- 46.Torres-Millan J, Torres-Lopez M, Benjumea-Serna M (2010) Location of the central venous catheter tip in the right atrium: description in 2348 critical patients. Med Int 34:595–599Google Scholar