INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Central venous catheterization of the subclavian or internal jugular veins is common in the intensive care unit (ICU). This procedure is followed by complications in 0.3 to 12% of cases according to the experience of physicians and to the definition of complications (1). Other than artery puncture, hematoma, vascular or neural injury related to the needle or to the guide wire (2), pneumothorax and catheter-tip misplacement can occur. The diagnosis of these two latter complications requires a chest radiograph (3, 4). Nevertheless, despite its simplicity, chest radiography may be time-consuming and the entire procedure (from the phone call to the radiology department to receiving chest radiography results) may require more than 30 min and could be harmful in the case of critically ill patients. Moreover, several investigators have questioned the need of postvenous cannulation chest radiography in the absence of clinical complications (5). Finally, in order to use hospital resources to the maximum and to minimize radiation, it could be useful to use alternative methods to conventional radiography (3).

Recent data have suggested that ultrasound can accurately detect pneumothorax in critically ill patients (6, 7). Furthermore, real-time echography is an easy technique to investigate the subclavian and internal jugular veins and can improve the success rate of catheter insertion (8). Last, whereas ultrasounds allow visualization of central venous catheters (12), this method has never been reported as a tool to detect catheterization complications.

We hypothesized that bedside ultrasonic examination performed by ICU physicians could accurately detect misplacement of the catheter and pneumothorax after catheterization of internal jugular and subclavian veins.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The study was conducted from March to September 1999 in a 14-bed medical ICU in a 821-bed tertiary care teaching hospital. The study was conducted according to the local ethics committee guidelines for human research. Informed consent was obtained from each patient or the next of kin.

Patients

All consecutive patients who underwent an internal or subclavian venous cannulation between 8:30 A.M. and 8:00 P.M. were evaluated. The patients included were ICU patients or patients from other wards. Cannulations that were performed other than during this period when only one physician was on duty were not included because of the blind design of the study. This procedure required two physicians: one who inserted the catheter and interpreted the chest radiograph and the other who performed the ultrasound examination and who was unaware of any cannulation procedure difficulties.

Procedure

Insertion site (jugular or subclavian) and side were chosen by the operating physician, based on the clinical situation and his personal preferences. The patient was placed in a supine position and catheters were placed under sterile conditions using the Seldinger technique. Dual-lumen 7-F, 16-cm catheters were used in the case of nontunneled catheterization (Multi-Med; Baxter Healthcare Corp., Irvine, CA). For tunneling, 53-cm Nutricath "Tunnels" catheters (Vygon, Ecouen, France) were inserted and secondarily cut. Ultrasounds were not used to localize or to identify vessels prior to catheter insertion.

After the insertion, the ultrasound examination was performed by another ICU physician. This examination was conducted as follows: (1) the examination of the two subclavian and internal jugular veins; in case of collapsed veins (especially in case of spontaneous breathing), the patient was asked to perform a Valsalva maneuver; (2) the visualization of the heart (right atrium and ventricle) and the inferior vena cava through the subcostal acoustic window; (3) pneumothorax detection using both lung-sliding and comet-tail artifact as previously described (6, 7). Briefly, examination of the lungs' ultrasound must identify the ribs. Between two ribs (identifiable by their acoustic shadow), the interface between the thoracic wall and the lung is a hyperechogenic line that represents both the visceral and the parietal pleura. Lung-sliding is a to and fro movement of this hyperechogenic line, which follows respiratory movements. Whereas the presence of lung-sliding rules out pneumothorax, its absence is associated with a pneumothorax in 90% of cases. Comet-tail artifacts are grossly vertical artifacts spreading from the hyperchogenic line to the inferior edge of the screen. The presence of comet-tail artifacts completely rules out pneumothorax, whereas their absence is specific to pneumothorax in less than 60%. Interestingly, it has been shown that the absence of both lung-sliding and comet-tail is in 96.5% cases specific to pneumothorax (7).

For the purpose of this study, pneumothorax detection was performed in the first three intercostal spaces. The presence of lung-sliding ruled out pneumothorax. Conversely, the absence of both lung-sliding and comet-tail-artifact meant pneumothorax.

We used a real-time ultrasound unit (Sonos 100; Hewlett Packard, Orsay, France) connected to a 3.5-MHz probe, except for vein visualization where a 7.5-MHz probe was used. Misplacement meant either an aberrant or a too distal position. An aberrant position was defined as catheter-tip visualization: (1) in the ipsilateral internal jugular vein or in the contralateral subclavian vein in case of subclavian puncture, or (2) in the ipsilateral subclavian vein or contralateral jugular vein in the case of internal jugular cannulation. Too distal a position was defined as visualization of the catheter tip in the intracardiac position intra-atrial or intraventricularor in the inferior vena cava.

Whereas the three investigators (E.M., J.G., and M.A.) did not have any training in ultrasonography, they had a 2-h specific training course, which was given by a radiologist. This training consisted exclusively of ultrasonographic practice. At the end of this training course, the three investigators performed three ultrasonographic examinations under the radiologist's control and afterwards were considered qualified to perform postprocedural ultrasonic examinations as described below.

As soon as the cannulation was performed, the radiology department was contacted, and the time delay in receiving the chest radiograph was measured. In the same manner, the entire time required to perform the ultrasonic examination (from moving the ultrasound unit until the examination was completed) was measured.

The answers to the three previously described items: existence of a pneumothorax, aberrant position, and too distal a position given by the ultrasonic examination and radiograph were compared. The postprocedural chest radiography was considered as the reference. As recommended, catheter-tip placement was considered as adequate when the catheter did not enter the atrium (13).

Statistical Analysis

Continuous variables were expressed as mean ± SD, and the time needed by the two methods was compared using Student's t test. Discrete variables were compared using the chi-square test. Statistical significance was set for a p value less than 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Eighty-five catheters inserted into 81 patients were included. The characteristics of these patients are given in Table 1. Among this population, 54 were hospitalized in the ICU, 21 of them received mechanical ventilation. Twenty-seven patients came from other medical wards. A subclavian approach was used for 70 catheters and an internal jugular approach for 15 catheters. Fourteen catheters inserted in non-ICU patients were tunneled. Six uncomplicated artery punctures were reported. Postprocedural chest radiography disclosed 10 misplacements and one pneumothorax (Table 2). Aberrant catheter position occurred in four cases: catheter tip in the ipsilateral jugular vein after cannulation of the left subclavian vein (n = 3), tip catheter in the right subclavian vein after cannulation of the right internal jugular vein (n = 1). Intracardiac position was observed in six cases: right atrium (n = 4), right ventricle (n = 2). The two intraventricular catheters were tunneled catheters. No catheter was observed within the inferior vena cava. Too distal a placement was more frequent among patients from other wards than among ICU patients, but this difference was not significant.

Among 255 examinations (vein examination: n = 85, pneumothorax research: n = 85, heart examination: n = 85) ultrasonic examination was interpretable in all cases except for one heart examination (feasibility, 99.6%). All too distal placements were recognized by echographic examination except in one severely obese patient in whom heart visualization was impossible. Nevertheless, this patient's ultrasonic examination ruled out aberrant position and pneumothorax.

Only one pneumothorax occurred after subclavian insertion and was identified by ultrasounds. This pneumothorax required chest tube drainage. No other pneumothoraces were diagnosed during the 48 h after insertion.

While performing heart and vein examination and pneumothorax research, ultrasound did not give any false-positive result.

The entire time required to perform ultrasonic examination was 6.8 ± 3.5 min, whereas it took 80.3 ± 66.7 min for the chest radiograph (p < 0.0001). In 27 cases, where radiography had been prescribed between 11:30 A.M. and 8:00 P.M., the time needed to get the chest radiograph was more than 100 min (maximum value, 300 min).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study, ultrasonic examination was highly accurate in detecting misplacement and pneumothorax after central venous catheter insertion. Insertion of central venous catheters in subclavian or internal jugular veins is quite frequent in the ICU. More than three million subclavian-vein catheters are placed in the United States each year (13). Even if this procedure is for most of the time uneventful, complications may occur, and even lead to severe distress (14) or death (15). In addition to evident difficulties such as the inability to locate or cannulate the vein, artery puncture, cervical hematoma or nerve injuries, catheter misplacement and pneumothorax are usually difficult to confirm in the absence of postprocedural chest radiography (3, 4). Although recent studies have underlined the cost of this procedure (3, 16, 17), Gladwin and colleagues (3) concluded that postprocedural chest radiography still remains necessary because clinical factors alone cannot reliably identify tip misplacement. Moreover, we observed in our unit that the time from the phone call to the radiology department to receiving the chest radiograph could be excessively long between 11:30 A.M. to 8:00 P.M.

These factors have led us to look for alternative methods to postprocedural chest radiograph. An accurate and prompt tool to detect postprocedural complications could be fluoroscopic examination, but this method implies a high dose of radiation and is rarely available except in the operating room and radiology departments. Furthermore, a fluoroscopic unit, which is more expensive than a real-time ultrasound unit, is used only for very limited indications and is generally cumbersome to move. Last, this technique cannot be utilized for all ICU beds. On the contrary, real-time ultrasonic examination is being used more and more frequently by ICU physicians to assess cardiac function or hemodynamic status (18, 19) and to diagnose pleural effusion, ascites, or obstructive renal insufficiency (20). These examinations can be performed with portable units. More recently, echography has been shown to be a very sensitive and specific method to eliminate pneumothoraces in critically ill patients (6, 7). The highly echogenic structure of modern intravenous catheter components makes the detection of these devices particularly easy in the main veins of the upper body and within the heart (12). Whereas bedside ultrasound guidance has previously been proposed to improve the venipuncture success rate during central venous catheterization (1, 8), to the best of our knowledge, bedside echography has never been used to assess misplacement and pneumothorax after this procedure.

Although ultrasonic examination was performed by physicians who did not have any training in ultrasound practice, this tool was highly accurate for the purpose of the study. In fact, our study suggests that ultrasonography practice when restricted to some specific and easy to analyze elements (visualization of the heart, of the jugular and subclavian veins, recognition of lung-sliding and comet-tail artifact) could be easily taught to ICU physicians.

As previously reported in similar conditions (3), we experienced a low rate of malpositions (12.9%). Whereas pneumothorax is a common complication of subclavian puncture, only one pneumothorax (1.2%) was observed despite a high percentage (82.3%) of subclavian puncture. This low rate agrees with the data of a previous large study (1). Only one pneumothorax occurred that was detected by ultrasonic examination. In our study, the lung-sliding and the comet-tail artifact recognition never needed more than 30 s.

Our work has suggested the feasibility and high accuracy of postprocedural ultrasonic examination to detect tip catheter aberrant position. Catheter placement in a main vein is an easy ultrasonic diagnosis. Optimal visualization of central veins by real-time echography can theoretically be impaired in hypovolemic or severely dehydrated spontaneously breathing patients, but this difficulty can be avoided by a Valsalva maneuver as previously reported (21). Conversely, among ventilated patients, central veins visualization is always easily obtained. As a result in our population, jugular and subclavian veins were always visualized perfectly.

Although transthoracic ultrasonography inability to visualize the heart is infrequent, examination of the heart was impossible in one patient whose catheter tip was intracardiac. However, in this patient, ultrasonic examination ruled out aberrant position and pneumothorax. This suggests that in obese or nonechogenic patients (COPD), ultrasounds could be inaccurate in detecting too distal a position.

Whereas real-time ultrasonography is safe in avoiding any radiation exposure, our study suggests that in addition this method saves time. This method could then be valuable in hemodynamically unstable patients who quickly need a well-placed catheter for the measurement of central venous pressure. Similarly, ultrasonic examination could quickly confirm pneumothorax and allow chest tube insertion in case of respiratory distress after catheter insertion. On the other hand, the time saved from ultrasonic examination could allow outpatients to be discharged more quickly from the ICU instead of waiting for the conventional chest radiography results.

In conclusion, this study suggests that ultrasonic examination is accurate in detecting pneumothorax and catheter misplacement after subclavian and internal jugular vein cannulation. Moreover, this procedure requires less time than conventional chest radiography and can be easily taught to ICU physicians after a 2-h training period.