Incidence of Posterior Vessel Wall Puncture During Ultrasound-guided Vessel Cannulation in a Simulated Model

Abstract

Objectives:  The incidence of posterior vessel wall puncture (PVWP) during central line placement with possible subsequent injury to structures lying behind the vein is unknown. At times the internal jugular vein lies immediately anterior to the carotid artery rather than lateral to it, leading to potential arterial puncture should the needle pass through the vein completely. The objective of this study is to evaluate the incidence of PVWP during simulated ultrasound (US)-guided vessel cannulation.

Methods:  Enrolled subjects were emergency medicine resident and attending physicians. Subjects performed US-guided venous access on simulated blood vessels within gelatin-based US phantoms. While blinded to the purpose of the study, each subject performed successful cannulation of the vessel on separate phantoms, with wire placement confirmed by expert review of a follow-up US. Each phantom was subsequently deconstructed to manually inspect for PVWP.

Results:  Thirty-five subjects with a range of experience in the technique participated, each performing both transverse and long-axis approaches for a total of 70 cannulations. The overall incidence of PVWP was 34% (95% confidence interval [CI] = 22.9% to 45.1%).

Conclusions:  This study found a high incidence of inadvertent PVWP during simulated US-guided vessel cannulation in this model.

ACADEMIC EMERGENCY MEDICINE 2010; 17:1138–1141 © 2010 by the Society for Academic Emergency Medicine

Arterial puncture during central venous catheter placement is a known complication of the procedure, with reported incidence up to 11%.^1,2 While ultrasound (US) guidance has been shown in several studies to decrease arterial puncture,^1–3 the identification of this event is based on the presence of forceful, pulsatile blood flow from the needle after removal of the syringe. Occult arterial injuries occur when the artery is punctured during the course of the procedure without resultant pulsatile flow through the needle.⁴ Both recognized and occult arterial injuries may result in significant complications.

When using a landmark technique for internal jugular venous access, many practitioners intentionally pass through-and-through the vessel and access the lumen when withdrawing the needle. The use of US allows for direct visualization of the internal jugular vein and internal carotid artery, but commonly demonstrates the vein superficial to the artery.⁵ In this scenario, any posterior vessel wall puncture (PVWP) that occurs could result in arterial injury. The objective of this study was to measure the incidence of PVWP during US-guided vessel cannulation using a phantom model.

Methods

Study Design

This was an observational study to assess the incidence of PVWP in a simulated model. The study was approved by the institutional review board.

Study Setting and Population

The study was conducted within an emergency medicine residency program with an active US curriculum. Study participants included physicians of various levels of training.

Study Protocol

Ultrasound-guided venous cannulations were attempted on phantoms with an internal simulated vein within a mixture of opaque gelatin as described in previous studies.^6,7 The simulated vein was made from multiple layers of commercially available plastic wrap melted together with direct heat around a cylindrical mold 8 mm in diameter. This size is comparable to the diameter of the adult internal jugular vein⁸ and the size used in commercially available central line simulators (Blue Phantom, Redmond, WA). The vein was filled with colored water, sealed at each end, and then examined for leaks. Each vein was filled with sufficient fluid so as not to collapse with gentle manual pressure. The veins were then positioned within the gelatin mold at a depth of approximately 1.5 cm. As part of the model design process, the study personnel constructed several of these phantoms and subsequently deconstructed them, ensuring that they could be pulled apart using only gloved hands (without the use of sharp instruments) and that the process of deconstructing left the vessel intact.

All participants received standard education on US-guided line placement consisting of a 1-hour lecture and training on commercially available simulators prior to involvement in this study. The approaches to short-axis and long-axis techniques were standard methods as previously described in the literature^6,9 and included emphasis on visualizing the needle tip during vascular access. Immediately prior to enrollment, each participant received an additional 15-minute didactic review on both techniques. Each study participant attempted two US-guided cannulations on two separate phantoms, one using a short-axis approach and one using a long-axis approach. US images were obtained using a 7.5-MHz linear array transducer and vascular presets (Sonosite MicroMaxx, Bothell, WA).

A successful cannulation of the vessel was defined as introduction of a standard 18-gauge needle into the vein, aspiration of fluid, and placement of a 0.89-mm spring guide wire. After the attempt, study personnel confirmed the wire location with an additional US of the phantom. Following each procedure, participants were asked if they had punctured the posterior wall during the attempt.

Each phantom was deconstructed and the veins were carefully removed and assessed for number of punctures. Additional fluid was added to the vessel as needed to create a slight positive pressure, enhancing the flow of fluid from punctures and ensuring that any leaks were easily visualized (Figure 1). Each phantom was labeled with a numbered code corresponding to the study participant and the approach used. Personnel removing the vessels were blinded to this information. Primary outcome measure was PVWP. Vessels with holes on opposite sides were classified as PVWP, whereas vessels with only a single hole were classified as no PVWP.

Figure 1.  The simulated vein. Left = intact vessel; middle = single vessel wall puncture; right = PVWP (through-and-through puncture; PVWP = posterior vessel wall puncture).

Data Analysis

Comparison of PVWP rates between short-axis and long-axis approach and between experience levels was performed using Fisher’s exact test. A post hoc power calculation revealed that with 35 subjects (70 attempts total), an alpha of 0.05, and beta 0.8, we were powered to detect a 25% absolute difference in percent of PVWP between groups.

Results

Thirty-five subjects with a range of previous experience in US-guided central venous access (range = 3 to >50) were enrolled. Twenty-three residents and 12 attending physicians participated. The average experience level of the physicians was 30 US-guided central lines performed prior to the study. Each subject performed two cannulations (one in short axis, one in long axis) for a total of 70 cannulations.

PVWP occurred in a total of 24 of the 70 attempts (34%, 95% confidence interval [CI] = 22.9% to 45.1%). The rate of PVWP in short axis for all operators was 31% (95% CI = 15.7% to 46.3%). The rate observed in long axis was 37% (95% CI = 21% to 53%), a difference that was not found to be statistically significant (p = 0.80).

There was no trend in rates of PVWP related to experience in US-guided central line placement. PVWP rates for operators with fewer than 10, 10–20, 20–30, 30–40, and more than 40 previous lines were 35, 31, 36, 50, and 38%, respectively (p = NS). There was, however, a greater rate of PVWP among residents (20 of 46 attempts, 43.5%, 95% CI = 29.2% to 57.8%) compared to attending physicians (4 of 24 attempts, 16.7%, 95% CI = 1.8% to 31.6%; p = 0.034). The difference in mean number of previous lines was not significant between residents and attending physicians (26.8 vs. 31.3; p = 0.60).

Nineteen of 24 (79%) cases of PVWP were unrecognized by the operator. Only one of the 46 attempts without a PVWP was incorrectly felt to have punctured the posterior wall.

Discussion

Ultrasound guidance decreases but does not eliminate the complication rate during central venous catheterization. In clinical studies with real-time visualization, the incidence of carotid injury during venous catheterization is as low as 1.1%.^1–4 The anteroposterior anatomical proximity of the carotid artery to the internal jugular vein may predispose it to inadvertent puncture.^1,5 The results of our study showing a high rate of PVWP provide preliminary evidence that in this anatomical scenario, the incidence of occult arterial injury could be much higher than previously reported.

Previous investigations into complications of central venous catheter placement such as arterial injury have been based on clinical evidence of these complications.^5,10 In contrast to previously published manuscripts, our results do not rely on indirect evidence of injury, but provide direct physical data on PVWP, albeit in a vascular phantom.

Using a long-axis approach during central venous catheterization has been suggested to decrease the incidence of PVWP by improving needle tip visualization.^1,9 In our study population, we were not able to detect a difference between PWVP in short- and long-axis approaches. There are two potential confounders that may have affected our results. While our study participants had been previously trained in both the long- and the short-axis techniques, and use both in clinical practice, we recognize that there is a preference among our physician staff with the short-axis technique. The relatively lower experience level in the long-axis technique may have resulted in the similar rate of PVWP. More importantly, our study was only powered to detect a difference of 25% between groups, and the actual difference may be smaller.

While experience level specific to US-guided lines did not correlate with different rates of PVWP, attending physicians showed a lower rate than residents. The fact that experience in US-guided venous access was similar in these two groups suggests that increased procedural experience in general serves to improve techniques specific to this procedure.

Limitations

This study was limited by small sample size. While we were able to assess the overall incidence of PVWP, our sample size was only powered to detect an absolute difference of 25% or larger between groups. Furthermore, small sample size prevented subgroup comparisons, such as short- versus long-axis technique in experienced versus inexperienced physicians. While the use of US phantom allowed for dissection and direct inspection of the vessel for injury, it does not replicate human tissue and therefore represents an additional limitation.

Finally, our study population included a wide range of experience levels that may have contributed to the increased PVWP rates. We contend that our study population reflects the range of experiences available in many emergency departments.

Conclusions

There was a high incidence of unrecognized puncture of the posterior vessel wall during ultrasound-guided venous cannulation of a phantom in this simulation model. These preliminary data suggest that this incidence is unchanged with different techniques, but decreases with higher levels of medical training. Further studies in larger populations are warranted before drawing definitive conclusions.