A randomised crossover study to compare the cross-sectional and longitudinal approaches to ultrasound-guided peripheral venepuncture in a model

Achieving intravenous (IV) access is one of the most common, and most important, clinical skills used in hospital medicine. In the majority of patients, such vascular access is provided by a peripheral intravenous (PIV) cannula. Traditionally, PIV access is achieved by a landmark-and-palpation approach; however, locating a suitable vein by this method is not always straightforward.

Finding and accessing a peripheral vessel in patients with a history of IV drug abuse, severe obesity or chronic medical problems can be impossible using the landmark-and-palpation approach alone. Failure rates for this skill range from 9 to 56% [1]. The alternative in such cases is commonly a central line insertion.

While effective, central venous access is not without risks and is often inappropriate for patients requiring only a short duration of IV treatment. Catheter-related sepsis, arterial puncture, pneumothorax and arrhythmias are all recognised complications of central venous access, and can cause significant morbidity and mortality for these patients [2].

The increasing portability of ultrasound machines has led to a surge in the availability of point-of-care ultrasound (PoCUS) in the hospital environment. Ultrasound-guided central venous access is now considered the standard of care, improving both success rates and reducing complications such as arterial puncture [3]. It has been suggested that the use of ultrasound to guide the cannulation of the deep brachial or basilic veins in patients with difficult peripheral access may improve success rates and, ultimately, patient outcomes [4, 5].

Ultrasound-guided peripheral vascular access is a safe alternative to central line placement and could be used instead in up to 85% of patients classified as having difficult IV access [6]. Training students and staff to locate and access deep peripheral veins under ultrasound guidance may therefore lead to fewer central line placements and, consequently, a reduced incidence of central line-related complications.

There are two commonly used approaches for ultrasound-guided PIV access: the cross-sectional approach (short-axis or out-of-plane approach) and the longitudinal approach (long-axis or in-plane approach). Previous studies have shown that it is possible to teach this skill to novice ultrasound users effectively, using a single 30-min didactic training session [7, 8]. It is unclear as to whether novices find it easier to learn and effectively perform the cross-sectional approach or the longitudinal approach, and there is no consensus which approach should be taught. We devised this trial to identify which approach (or approaches) ought to be taught to novices and to help inform the development of a structured learning package.

We chose to conduct a prospective randomised controlled trial using medical students in order to answer this question. It has been demonstrated that previously ultrasound-naïve medical students are able to use ultrasound to guide peripheral vascular access on real patients after a single training session [9]. It has also been demonstrated that simulation-based training using a dedicated ultrasound phantom can lead to improved levels of skill when applied to real clinical practice [10]. However, we appreciate that we did not follow up phantom training with testing on patients. As such, we advocate testing with patients in future.

As we were testing the efficacy of novices, we used phantom models to simulate patients. Due to the construction of these phantom models, cannula lines tended to kink when inserted; thus, failure reflected the model limitations rather than the skill of the novice. To overcome this, we instead aimed for venepuncture, with the students required to aspirate 1 ml of fluid. This gave the benefit of requiring the students to correctly identify, locate and puncture the vein, while removing any failure bias due to the firm material of the model. We feel that as the main purpose of ultrasound guidance is to aid in locating and puncturing the vein, venepuncture acted as an adequate proxy for peripheral venous access.

The data for all 30 subjects were analysed. No student was excluded from the data analysis.

The mean time from first puncture of the model’s skin to successful aspiration using the cross-sectional approach was 45.1 s (95% CI 29.4 to 0.9 s). The mean time required using the longitudinal approach was 52.8 s (95% CI 36.7 to 69.0). Performing a paired samples t test, we found that the difference of 7.7 s (95% CI −12.1 to 27.4) was not statistically significant (p = 0.43).

The mean number of skin punctures required using the cross-sectional approach was 1.13 (95% CI 0.94 to 1.32). The mean number of skin punctures required with the longitudinal approach was 1.30 (95% CI 1.08 to 1.52). Paired sample t-tests demonstrated that this difference of 0.17 skin punctures (95% CI −0.08 to 0.41) was not statistically significant (p = 0.17).

The mean perceived difficulty score reported for the cross-sectional approach was 3.97 out of 10 (95% CI 3.10 to 4.83). The mean perceived difficulty for the longitudinal approach was 3.93 out of 10 (95% CI 3.21 to 4.65). The difference in mean perceived difficulty of 0.04 out of 10 (95% CI −0.95 to 1.01) was not statistically significant (p = 0.95).

For graphical representations of the results, see Figs. 3, 4 and 5.

Since the first emergency department observational study of the use of ultrasound to establish peripheral vascular access in 1999, [5] the availability of both ultrasound equipment and the ultrasound expertise of emergency physicians has expanded considerably. In many emergency departments, ultrasound is being increasingly utilised to obtain PIV access in patients where attempts using the traditional landmark-and-palpation technique have failed [2].

Ultrasound-guided PIV cannulation is also being utilised in other settings such as intensive care units [13]. Use of ultrasound provides a very real alternative to more invasive approaches such as central venous access, [6, 13] external jugular vein, [14] venous cutdown or intraosseous access. In addition, ultrasound-guided PIV cannulation has been shown not to pose a significant infection risk [15].

With regard to the efficacy of ultrasound-guided PIV cannulation over a traditional landmark-and-palpation technique, the literature is conflicting. Whist Constantino found that ultrasound was more successful than traditional techniques [16], not all the literature is positive. Several studies have failed to show any advantage of ultrasound-guided peripheral vascular access over a traditional technique [17, 18]. A study in children under 10 years also failed to show any statistical advantage of ultrasound over a traditional approach [19].

Although ultrasound-guided PIV cannulation has not been shown to be better than traditional techniques in all patients, there is evidence to support its use in difficult patients or where traditional techniques have failed. The international evidence-based recommendations on ultrasound-guided vascular access from 2013 recommended that, “Use of ultrasound should be taken into consideration for any kind of peripheral intravenous line when difficult access is anticipated” [20]. They also recommend the use of the longitudinal approach but accept that it is more challenging than the cross-sectional approach. The longitudinal approach may be safer by allowing better needle tip visualization and thereby minimising posterior wall puncture [8]. With ultrasound-guided central venous access, it can be difficult to distinguish between the vein and the artery in longitudinal view, with a greater risk of arterial puncture with the longitudinal approach. The cross-sectional approach allows identification of not just the vessel to be punctured (the vein) but also the vessel to be avoided (the artery). This visualisation of both vessels is lost with a longitudinal approach.

There is currently no consensus as to whether a longitudinal or a cross-sectional approach is superior. In a single study Mahler showed that the cross-sectional approach was slightly faster, but failed to demonstrate any further benefit over the longitudinal approach [21].

How to teach ultrasound-guided PIV cannulation is less well covered in the literature. After a short training package, perceived difficulty of obtaining peripheral vascular access was reduced in a cohort of emergency nurses [4]. A study involving utilisation of ultrasound to aid PIV access learning by medical students using just the short-axis approach showed no difference in terms of obtaining PIV access, but the ultrasound group did perceive the experience as easier and felt they had gained more knowledge of the mechanics of placing an IV cannula [9].

There has been some focus on how best to teach this skill and also which approach is best (cross-sectional versus longitudinal). Blaivas et al. [7] found a clear superiority of the cross-sectional approach over the longitudinal approach in terms of time taken from placing the ultrasound probe and needle to successful vascular access. Interestingly, our study generated some quite different results, and failed to show a difference despite our sample size being almost twice that used by Blaivas and colleagues. (30 subjects, compared to 17 subjects).

Our study has a number of limitations that should be acknowledged. Firstly, our subjects were required to access simulated vessels in an ultrasound phantom, as opposed to the real vessels in a living human patient. The task of aspirating fluid from a model under ultrasound guidance is undoubtedly easier than taking real blood from a patient, as our subjects did not have to attend to the model’s concerns or discomfort during the procedure. Additionally, the deep peripheral veins in real patients are unlikely to be as straight as the vessels in our inanimate phantom. Acquiring and maintaining an adequate longitudinal view of a tortuous vein in a real patient is likely to be considerably more difficult than obtaining a good longitudinal view in our model. As such it is difficult to extrapolate our findings to clinical practice, and this study does not provide evidence of benefit of one approach over the other in clinical practice. The findings are limited to the training of novices using a phantom model.

It should also be noted that our study measured only whether our subjects could achieve venepuncture, while in reality the insertion of a PIV cannula requires an additional step in which the catheter is inserted into the vessel following venepuncture. We appreciate that venepuncture does not include all the difficulties of cannulation, for example, confirmation of patency and the increased risk of puncturing of the posterior wall. However, as the purpose of ultrasound is to guide the user in locating the vessel rather than the technique of cannulation, we considered it an adequate proxy.

This study has demonstrated that, in novice ultrasound users, neither the longitudinal nor transverse approach showed a clear superiority when ultrasound-guided venepuncture was attempted in a model. More research is needed to definitively answer the question of whether the cross-sectional or longitudinal approach is more effective when teaching this skill to novice ultrasound users. Future researchers may wish to have subjects attempt the technique on human subjects to produce results that translate clearly to real clinical practice. However, considering the clear benefits of peripheral venous access (as demonstrated in other studies), the most pertinent point may be to increase teaching of ultrasound-guided peripheral venous access in general, regardless of approach [2, 3]. Until such times as this question is resolved, we believe that both the longitudinal and cross-sectional approaches have value, and that both techniques should be included when developing a training programme to teach this useful clinical skill.