In teaching hospitals in many countries, central vein cannulation (central venous catheters (CVC) or dialysis catheters) is usually performed by residents. Whenever possible, the real-time ultrasound-guided (UG) technique has become the recommended procedure for central vein catheterization in ICU and emergencies, because it increases the success rate, and decreases the complication rate, the procedure time and the cost 1 2 3 4 5 . According to guidelines, novice residents should start by learning the UG technique. However, the American Society of Echocardiography, the Society of Cardiovascular Anesthesiologists, the National Institute for Clinical Excellence and the American Society of Anesthesiologists all recognize that, in certain circumstances, such as emergency situations, the use of ultrasound may be impossible (because unavailable) and that operators must therefore maintain their skills by placing central catheters according to the landmark (LM) technique for these specific situations 1 2 3 . Some authors have suggested that the anatomical knowledge gained by using the ultrasound technique improves the operators’ skills when they need to use the LM technique in an emergency 6 . However, this opinion is based exclusively on personal experience and not on any clinical data. It could also be harmful for physicians who have only learned the UG technique to consider themselves sufficiently skilled to attempt an LM procedure in an emergency situation. In our experience, residents performing only UG catheterization are not able to achieve the LM technique and therefore require additional training. In our ICU, inexperienced residents exclusively learn the UG technique during the first 3 months of their rotation and are then allowed to learn the LM technique during the following 3 months under the supervision of the attending physician. The objective of this study was to determine the learning curve of the LM technique in residents who have only been trained in the UG technique.

In accordance with French legislation, the local institutional review board (CPP Nord-Ouest II, Amiens University Hospital, France) approved the study protocol and considered no written informed consent was required, as teaching the LM technique is recommended and part of our regular practice. Information was given and oral consent obtained from patients or their relatives about the purpose of this prospective observational study and the anonymous use of the parameter recorded for scientific publication. Over a 3-year period (May 2008 to April 2011), all residents with no experience of catheter placement (fewer than three attempts) working in our medical ICU (Amiens University Hospital, France) were included in this study. All residents included in the study were followed during their 6-month rotation in our unit. During the first 3 months, the residents were only allowed to use the UG technique. During the following 3 months, residents were allowed to perform CVC placement by the LM technique under the supervision of the attending physician. The decision to allow the resident to use the LM procedure was based on the risk of complications and the patient’s condition, as assessed by the attending physician. All CVC or dialysis catheters, femoral or jugular procedures performed by these residents were prospectively recorded during this 6-month period (except for UG procedures performed during the LM period).

Subclavian catheter placement is rarely performed in our department and never by trainee residents because of the high risk of pleural puncture, and because many patients in our ICU (part of the nephrology department) require careful preservation of their capital vein in case a fistula is subsequently required. The central catheter placement-learning program in our department was established according to guidelines 1 . Before starting their first procedure, all residents received theoretical training on ultrasound, the ultrasound apparatus and central line placement using the UG technique. For that purpose they also watched the video provided by the New England Journal of Medicine named ‘Placement of a femoral venous catheterʼ, which is available online. Because we do not have any inanimate models, the residents learned on patients how to manipulate the probe to correctly visualize the different vessels, and then performed a four-hands procedure with the attending physician. After this educational training program (including the four-hands procedure), they were allowed to perform their first procedure. At the beginning of the fourth month, when residents started to perform LM procedures, they received an additional tutorial on the anterior LM technique. All UG and LM procedures were performed under the direct supervision of a senior physician.

The UG technique has been shown to improve the safety and efficacy of CVC placement in ICU patients compared to the LM technique and is now the most broadly recommended procedure 1 2 3 4 5 . However, in some emergency situations, an ultrasound device may not be available. Physicians must therefore be able to perform CVC placement according to the LM technique. It is unclear whether residents who have only learned the UG technique are also able to perform the LM technique. The present study clearly demonstrates that residents who had only learned the UG technique were not immediately able to perform the LM technique, but required specific training in this technique, comprising at least 10 LM procedures.

Previous studies by our team and other authors have already demonstrated the benefits of the UG technique for physicians inexperienced in CVC placement, especially during the first procedures when UG can markedly decrease the complication rate 8 9 10 . The UG learning curve is steeper than the LM learning curve. In other words, residents achieve success earlier with the UG technique with fewer complications. Therefore, by exclusively teaching the UG technique, patients would be managed by the safer of the two procedures. However, such an approach means that young physicians may be unable to perform catheter placement without ultrasound guidance, especially in emergency situations.

This study, confined to novice residents, addressed this issue and clearly showed that training exclusively in the UG technique does not provide sufficient skills to perform the LM technique. This study also showed that the residents’ skills rapidly improve after they have performed 10 procedures with a success rate (81%) and complication rate (6%) comparable to those observed in previous studies of physicians learning the LM technique 7 10 .

These results also demonstrate that learning the UG technique provides residents with certain skills to perform LM catheterization, as first the residents only needed to perform 10 LM procedures to achieve satisfactory success and complication rates and second the procedure time was not increased when residents used the LM technique although venous return time increased. Use of UG helps the physician to puncture the central vein. The venous return time is therefore prolonged when switching to the LM technique, but once the needle is in the vein, the catheter insertion is the same using the UG technique and the LM technique. Figure 2 also shows that the procedure time, which includes venous return time and catheter insertion, improved continuously with the number of procedures, even when the resident switched to the LM technique, demonstrating that the resident acquires certain skills during UG procedures that are useful for central catheter insertion.

This acquisition of skills also explains why a shorter procedure time was observed in the LM group. In our study, UG procedures were performed before LM procedures, so residents had already acquired a certain amount of experience when they started to perform LM procedures. In previously published studies that demonstrated a gain of time with UG compared to LM, the two techniques were randomly assigned avoiding any difference of experience between groups 3 8 .

This study shows the need to teach the LM technique to residents already skilled in the UG technique so that they will be able to perform catheter placement when an ultrasound machine is not immediately available. However, our study also demonstrates that, even after the resident has acquired good skills with the UG technique, the learning phase of the LM technique is still associated with significant complication and failure rates. The results of this study therefore raise a more general issue. Instead of teaching residents to use the LM technique when ultrasound is not available, which exposes patients to a risk of complications, alternative emergency approaches, such as intra-osseous lines could be considered 11 . The use of simulators, which have already been shown to improve physician training in cannulation techniques, could also be a safer alternative to teach the LM technique 12 13 . Future investigations should address these questions raised by our study.

We did not address the issue of maintenance of these LM procedure skills, notably whether experienced physicians should occasionally perform the LM technique in order to maintain their skill. This study concerned various sites (femoral and jugular) and types of catheterization (CVC or dialysis catheter). The study could not have been designed otherwise, as residents are trained simultaneously in jugular and femoral vein catheterization and the learning curve depends on each procedure regardless of the site. During the LM technique period, the attending physician was able to oblige the resident to use the UG technique. This decision was based on evaluation of the risk factors and may have constituted a selection bias in favor of the LM group, as patients in the LM group would have presented fewer risk factors for complications. However, the two groups presented similar risk factors (Table 1). The only difference between the two groups was a higher proportion of femoral procedures in the LM group. Success rates (72% versus 71%, respectively; P = 0.9) and complication rates (22% versus 21%; P = 0.9, respectively) were similar between jugular and femoral procedures in the LM group (Table 2). We therefore consider that this selection did not constitute a significant bias in our study. Our LM procedure uses the thyroid cartilage as the palpation landmark, whereas other teams recommend using the cricoid cartilage. To the best of our knowledge, no data are available in the literature in favor of the use of one or other of these cartilages and we consider the thyroid cartilage to be more easily palpable. Our results in the UG procedure were obtained using an ultrasound machine with a needle guide attached to the transducer, a device specifically dedicated for the placement of CVC 7 .

The UG technique is the first technique that should be taught to novices, as it is associated with a steeper learning curve and a lower complication rate for patients. However, in order to ensure that residents are adequately skilled in all situations, especially when an ultrasound machine cannot be used in an extreme emergency, they must be able to perform catheter placement by the LM technique. This study demonstrates that training in the UG technique provides the resident with certain, but insufficient skills for catheter placement by the LM technique. Residents still require a training program comprising at least 10 LM procedures to achieve optimal skills. In view of the complication and failure rates associated with these procedures, it is unclear whether or not the LM technique should still be taught to novice residents and which alternative method could be use in emergency situations when an ultrasound machine is not available.

• The real-time ultrasound-guided technique is the recommended procedure for central vein catheterization. However, in emergency situations an ultrasound machine may be unavailable.

• To ensure that physicians are adequately skilled in all situations, they must be able to perform catheter placement without ultrasound.

• Training in the ultrasound-guided technique provides the resident with certain, but insufficient skills for catheter placement by the landmark technique.

• A training program comprising at least 10 landmark procedures is required to achieve optimal skills.

BMI: body mass index; CVC: central venous catheter; DBP: diastolic blood pressure; GEE: generalized estimating equations procedure; LM: landmark; SAPS2: simplified acute physiologic score 2; SBP: systolic blood pressure; UG: ultrasound-guided.

All authors declare that they have no competing interests.

JM, LG, PH, SSM and BC contributed to the acquisition and analysis of data. JM drafted the manuscript. JM, GC and MS contributed to the conception and design of the study. JM, LG, PH, SSM, BC, GC and MS were involved in the interpretation of data and critically revising the manuscript. All authors read and approved the final manuscript.