|
|
| |||||||||||||||||||||||||||||||||||||||
Background Although catheterization of the subclavian vein is a common procedure, the risk factors for complications and failures, with the exception of the physician's experience, are poorly understood. Ultrasonography has been recommended to help guide the placement of central venous catheters.
Methods We conducted a prospective randomized trial of ultrasound-guided location of the subclavian vein as compared with standard insertion procedures. In the group of patients undergoing catheterization with ultrasound guidance, the site of the insertion was marked before the catheterization attempt; real-time ultrasound guidance was not used. The 821 eligible patients (411 in the ultrasound group and 410 in the control group) underwent catheterization in a single procedure suite under controlled nonemergency conditions, in most cases for the administration of chemotherapy.
Results Ultrasound guidance had no effect on the rate of complications or failures of subclavian-vein catheterization (risk ratio for complications, 1.00; 95 percent confidence interval, 0.66 to 1.52; risk ratio for failures, 1.04; 95 percent confidence interval, 0.72 to 1.50). In multivariate analyses, prior major surgery in the region (P = 0.002), a body-mass index (the weight in kilograms divided by the square of the height in meters) higher than 30 or lower than 20 (P = 0.009), and previous catheterization (P = 0.043) were associated with failed attempts. Complications were also associated with failed attempts: 52 of the 721 patients (7.2 percent) in whom catheterization was successful had complications, as compared with 28 of the 100 patients (28 percent) in whom physicians were unable to place catheters. The number of needle passes was strongly associated with the rates of failure and complications. The complication rate rose from 4.3 percent with one pass to 24.0 percent with more than two passes.
Conclusions Ultrasound guidance of subclavian-vein catheterization, as used in this study, was not beneficial. In patients at highest risk for complications and failures, catheterization should be attempted by the most experienced physicians available.
Subclavian venipuncture is often a successful and uncomplicated procedure. Reported complication rates range from 0.3 to 12 percent, according to the experience of the physician and the definition of complications.1,2,3,4,5,6,7,8,9 Potential complications include failure to locate or cannulate the vein, puncture of the subclavian artery, misplacement of the catheter (defined as placement of the catheter tip in the contralateral subclavian vein or in either jugular vein), pneumothorax, mediastinal hematoma, hemothorax, and injury to adjacent nerves. With the exception of the physician's experience, the risk factors for complications and failures of subclavian-vein catheterization are poorly understood.
Ultrasound has been recommended to determine the position of the vein or to guide physicians in placing central venous catheters.10,11,12,13,14 We evaluated the factors associated with complications and failures of subclavian-vein catheterization and determined whether ultrasound guidance in locating the subclavian vein could minimize complications and failures.
Methods
A prospective randomized trial of ultrasound guidance in locating the subclavian vein, as compared with the standard insertion procedure, was performed at the University of Texas M.D. Anderson Cancer Center with the approval of the institutional review board. All patients participating in the study provided written informed consent. The goal was to enroll 1100 patients, but the study was closed after 824 patients had been enrolled, because an interim analysis showed that ultrasound guidance as performed in the study had no effect on the success of catheterization.
The following factors were evaluated: age, sex, height, and body-mass index (the weight in kilograms divided by the square of the height in meters); the time from the initial insertion of the catheter until it was placed; the number of years of postgraduate training of the physician inserting the catheter; the number of needle passes (defined as separate skin punctures) attempted; the side on which the catheter was inserted; prior catheterization or attempted catheterization on the same side; and prior major surgery (e.g., mastectomy, neck dissection, axillary dissection, or thoracotomy) or radiotherapy in the same region. All procedures were monitored by a research nurse for arterial punctures, nerve injuries, and the number of passes.
A chest film obtained after the procedure was evaluated for the presence of pneumothorax, hemothorax, mediastinal hematoma, and misplacement of the catheter. If the first physician attempting to insert the catheter failed to do so, this attempt was defined as a failure for the purpose of the study. In most patients in whom the first attempt at catheterization failed, another attempt was made by a second physician. The outcome of these subsequent attempts was also recorded. If a catheter was misplaced, the position was corrected either by a "power flush" (a rapid infusion of 10 ml of saline solution pushed through the catheter with a syringe) or by manipulation of the catheter under fluoroscopic guidance by a radiologist. Pneumothorax was treated with tube thoracostomy if it was symptomatic or progressive or if more than 20 percent of the interface between the lung and the chest wall was separated. Block randomization was used to ensure equal numbers of patients in the group undergoing catheterization with ultrasound guidance and the group undergoing catheterization with the standard procedure for insertion.15
Procedures for Insertion
For standard insertion of the catheter, the patient was placed in a supine position. The ipsilateral anterior superior region of the chest was shaved if necessary, degreased with acetone, and prepared in a sterile fashion with povidone-iodine. The patient then was placed in Trendelenburg's position. Lidocaine solution (1 percent) was injected into the infraclavicular puncture site with a 22-gauge needle. Physicians were encouraged to locate the subclavian vein with this 22-gauge needle; however, this 3.81-cm needle is often not long enough to reach the vein. The subclavian vein was punctured with an 18-gauge needle. The return of venous blood into a syringe attached to the needle confirmed entry into the vein. A guide wire was passed through the needle into the vein, and the needle was removed. A dilator was passed over the wire and removed. For single-lumen cannulation, a 5-French Silastic catheter (Davol, Salt Lake City) was passed over the wire, the wire removed, and the return of venous blood through the catheter confirmed. Dual-lumen 7-French Silastic catheters (Davol) were inserted through a peel-away sheath introducer.
For catheterization with ultrasound guidance, a 7.5-MHz linear-array ultrasound probe connected to a real-time ultrasound unit (Model 633, Aloka, Wallingford, Conn.) was used to measure the depth and caliber of the subclavian vein at the level of the clavicle, evaluate its patency, and mark its location on the skin. The standard procedure was then followed to prepare the skin and insert the catheter. Real-time ultrasound guidance was not used for placement of the catheter.
The physicians inserting the catheters ranged from surgical interns to surgical staff with 12 years of postgraduate training. Forty-nine physicians participated in the study, each inserting 1 to 62 catheters. At our institution, 8 to 20 patients per day undergo subclavian-vein catheterization in a single procedure suite, in most cases for the administration of chemotherapy. The majority of patients in this study had breast cancer or lymphoma. All the catheters were inserted under controlled, nonemergency conditions.
Statistical Analysis
The statistical analysis included logistic regression and likelihood-ratio chi-square tests. For each variable, differences between the two groups were initially evaluated with a univariate logistic-regression model that contained a single independent variable. Significance was determined according to the P value of the likelihood-ratio test. A forward stepwise method of selection was used to construct models that identified the most important risk factors and predicted the probability of a successful procedure. The likelihood-ratio chi-square test (performed with contingency tables) was used to determine the factors that affected success. This test is equivalent to the likelihood test for the significance of the coefficient for a variable in a univariate logistic-regression model.16,17 A P value less than 0.05 (two-tailed test) was considered to indicate statistical significance. Relative risk was used to determine the effect of ultrasound guidance on the placement of the catheter.18
Results
A total of 824 patients were eligible for the study; 3 were excluded because they were unable to maintain a supine position, which precluded the attempt to insert a catheter. Of the 821 patients, 411 were randomly assigned to undergo catheterization with ultrasound guidance, and 410 were assigned to undergo catheterization with the standard insertion procedure (the control group). A total of 468 patients (57 percent) were women, and 353 (43 percent) were men; 208 (25 percent) of the catheterization attempts were made on the left side, and 613 (75 percent) were made on the right. Complications occurred in 40 patients in the ultrasound group (9.7 percent) and in 40 in the control group (9.8 percent). There were 51 failed attempts at cannulation in the ultrasound group (12.4 percent) and 49 (12 percent) in the control group. In the ultrasound group, as compared with the control group, the risk ratio for complications was 1.00 (95 percent confidence interval, 0.66 to 1.52), and the risk ratio for failure was 1.04 (95 percent confidence interval, 0.72 to 1.50). Because the technique of ultrasound guidance we used had no effect on the frequency of complications or failures, all patients in the study were considered together in the evaluation of other risk factors for adverse outcomes. Failed attempts at cannulation were considered separately from complications. The risk factors for complications and failures are shown in Table 1.
|
Initial attempts at cannulation failed in 100 patients. In the univariate analysis, the factors associated with failure were previous major surgery or radiation therapy in the region, prior catheterization, prior attempts at catheterization, only one year of postgraduate training on the part of the surgeon performing the procedure, a high body-mass index, and more than two needle passes. Among the patients with a body-mass index higher than 30, the failure rate was 20.1 percent. If only one needle pass was attempted, the failure rate was 1.6 percent, as compared with 10.2 percent for two passes and 43.2 percent for three or more passes. In the multivariate analysis of preexisting risk factors, prior surgery (P=0.002), a body-mass index higher than 30 or lower than 20 (P=0.009), and prior catheterization (P=0.043) were associated with failed attempts. Combinations of risk factors were associated with higher failure rates (Table 2). There were 22 failed attempts (8.1 percent) among the 272 patients who had a body-mass index between 20 and 30 and no prior surgery or catheterization, as compared with 78 failed attempts (14.2 percent) among the 549 patients with prior surgery or catheterization or a body-mass index lower than 20 or higher than 30 (P<0.001).
|
Eighty patients (9.7 percent) had complications. These included misplacement in 49 (6 percent), arterial puncture in 30 (3.7 percent), pneumothorax in 12 (1.5 percent), and mediastinal hematoma in 5 (0.6 percent). Sixteen patients (1.9 percent) had more than one complication.
In the univariate analysis, the patient's sex, the body-mass index, and the number of needle passes were associated with the rate of complications. Women were more likely than men to have complications (11.8 percent vs. 7.1 percent, P=0.03). A body-mass index lower than 20 was associated with a complication rate of 16.2 percent, as compared with a 10.4 percent rate among patients with a body-mass index between 20 and 30 and a 4.0 percent rate among those with a body-mass index higher than 30. The complication rate was 4.3 percent for one needle pass, 10.9 percent for two needle passes, and 24.0 percent for three or more passes. No association was found between the complication rate and the years of postgraduate training on the part of the physician performing the procedure. In the multivariate analysis only a body-mass index below 20 was associated with complications (P=0.019).
The strongest predictor of a complication was a failed catheterization attempt. Of the 721 patients in whom cannulation was successful, 52 (7.2 percent) had complications; in 33 of these patients (63.5 percent) the complication was misplacement of the catheter. Of the 100 patients in whom physicians were unable to place catheters, 28 (28 percent) had complications. By definition, none of these complications included misplacement of the catheter.
Outcome after Failure of the Initial Cannulation Attempt
In 87 of the 100 patients in whom the initial attempt at catheterization failed, a subsequent attempt was made by a second physician. Catheterization was successful in 80 (92 percent) of these patients. In 64 patients (73.6 percent) the catheter was inserted on the first or second attempt. This success rate was similar to that for the initial attempt with one or two passes (successful insertion in 638 of the 821 patients, or 77.7 percent). There were complications in 7 of the 87 patients (8 percent) in whom a second physician attempted to insert a catheter. A third physician attempted catheterization in two patients; in both instances, the first needle pass was successful.
Discussion
We found no benefit associated with ultrasound guidance in locating the subclavian vein for catheterization. There are several possible reasons. First, many different physicians inserted the catheters. Studies reporting favorable results with the use of ultrasound guidance have typically involved only a small number of physicians. Second, physicians may become more proficient in the placement of catheters with ultrasound guidance as they gain more experience with this approach; however, we noted no such trend. Third, real-time ultrasound guidance of catheter placement, which may be beneficial in difficult cases, was not used in our study.
Some risk factors for failures and complications can be identified before the insertion of the catheter, whereas others may be apparent only during catheterization. For patients with a body-mass index higher than 30 or lower than 20 or with a history of previous catheterization or prior surgery or radiotherapy on the side selected for catheterization, the best approach may be to have experienced physicians attempt catheter placement rather than physicians who are learning the procedure.
Conventional wisdom suggests that a subclavian catheter not be inserted on the side of the body where an axillary dissection has previously been performed because of an increased risk of edema or infection. Our findings suggest another reason: a substantially higher failure rate for catheterization on the same side. Prior surgery or radiotherapy may cause a slight shifting of the position of the subclavian vein or an alteration of the surface landmarks used to locate the vein. When possible, the contralateral side should be selected for catheterization in patients who have undergone previous catheterization, surgery, or radiotherapy in the clavicular region.
The explanation for the higher frequency of complications in women than in men is unclear, but it may be related to differences in body habitus. The predominant selection of the right side for catheter placement in this study reflects the belief at our institution that the rate of thrombosis associated with catheterization is lower on the right side than on the left.
Given the increased likelihood of failure and the markedly increased risk of a complication when more than two needle passes were attempted by the same physician, we believe that more than two attempts by the same physician should be discouraged, particularly when catheterization is elective. We found that when one physician failed to place the catheter, the success rate for an attempt by a second physician (whose degree of experience was similar to that of the first) with one or two needle passes was similar to the success rate for the cohort as a whole on the first attempt. The complication rate was also similar.
Supported by the Office of the Vice-President for Patient Care, University of Texas M.D. Anderson Cancer Center.
Dr. Hohn designed the single- and double-lumen catheters used in this study and receives royalties from Davol, Inc., the manufacturer.
We are indebted to Dr. Glen Isaacson for his invaluable insights; to Drs. Charles Balch and James Cox for institutional support; to Anita Aguerre and Charlene Waldron, who performed the ultrasound examinations; to Ines Spencer and Bonnie Souders, who provided support as research nurses; and to the M.D. Anderson Infusion Therapy Team.
Source Information
From the Departments of Surgical Oncology (P.F.M., D.C.H., D.M.O.), Diagnostic Radiology (B.D.F.), and Academic and Research Computing (M.A.G.), University of Texas M.D. Anderson Cancer Center, Houston.
Address reprint requests to Dr. Mansfield at the University of Texas M.D. Anderson Cancer Center, Department of Surgical Oncology, Box 106, 1515 Holcombe Blvd., Houston, TX 77030.
References
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
Complications and Failures of Subclavian-Vein Catheterization
Jobes D. R., Ellison N., Troianos C. A., Weber M., Huber C., Oates A., Hajela V., Channer K.S., Kaufman J. L., Eynon C.A., Johnson J. R., Mansfield P. F., Gregurich M. A., Haire W. D.
Extract |
Full Text
N Engl J Med 1995;
332:1579-1581, Jun 8, 1995.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
