Background:

The current state of treating flow limiting coronary artery lesions is to perform manual percutaneous coronary intervention (mPCI). In the last decade there has been immense effort to standardize this procedure by implementing a robotic angioplasty tool to substitute the manual action, which was previously required. This led to a measurable decrease in operators health hazards when the procedure showed to be less time effective compared to mPCI. 

Purpose:

Since any new technique is associated with a specific training effect, the aim of our study was to analyse the learning effect with a second generation rPCI.

Methods:

All patients who underwent rPCI two centres (Giessen and Hamburg) were enrolled in this retrospective-pooled two-centre analysis. Primary endpoint of interest was the total procedure time (from the beginning of sterile coverage to the exit of the guiding catheter). Secondary endpoints were the volume of contrast agent, radiation data, as well as wiring time. To analyse the training effect curve fitting regression and residue analysis was performed. Next the turning point of best fitting regression was defined as end of the learning. Then the different groups were compared. 

Results:

A total of 104 patient datasets were included in this study. The baseline characteristics were consistent with a high cardiovascular risk typical of this collective with good angiographic outcome and no serious adverse events. In each regression analysis, only the total procedure time was associated with a significant regression model. 

The best fitting was a logarithmic (p = 0.038) as well as a cubic term (p = 0.051) with acceptable residue plots. A turning point of the logarithmic model was documented after the very first procedures, however the cubic model shows a learning curve after 15 procedures (in each department). After comparison of the two groups a significant difference between the early (n= 30; total time: 69.5 ± 42.2 min) and late group (n= 74; total time: 52.1 ± 31.8 min) could be documented. When comparing the baseline and procedural parameter, within the early group there were more complex procedures, an increase in contrast agent, and in fluoroscopy time. 

Conclusions:

We report preliminary data concerning the clinical application of rPCI systems. Our data suggesting a learning curve of 15 procedures. Nevertheless, the learning with the second generation rPCI system seems to be less department and more operator depended. Data on an operator basis are still missing. Further studies should address the training performance on an individual operator level within controlled conditions. 

Finally, rPCI is a new and interesting tool which now is utilised in clinical routine. Regardless of the unknown benefit for the patient, it offers the potential to reduce the health hazards for the operators, while there are no indications of inferiority in patient’s outcome.