A COMPUTER MODEL OF LARYNGOSCOPY AND LARYNGOSCOPE BLADE DESIGN
John H. P. Friesen MD
Department of Anesthesia, Victoria General Hospital, 2340 Pembina Highway, Winnipeg, Manitoba R3T 2E8
INTRODUCTION
Numerous different laryngoscope blades are available, all intended to make endotracheal intubation easier and safer. To discover better blade shape and design features, a mathematical model of direct laryngoscopy was constructed and implemented in the C++ computer language. A quantitative evaluation function was developed for use with an optimization algorithm.
METHODS
Blade shape was modeled by a curve (of N+1 line segments) calculated using an algorithm taking N independent input variables. The oral and pharyngeal anatomy was also modeled, using three parameters (mouth opening, inter-incisor angle, and relative position of the mandible). For each insertion depth (hyoid to incisor distance), the forward space and eyeline deviation (modified from indices proposed by Marks et al.1) were determined and used to calculate an evaluation function. For each different choice of anatomical parameters and insertion range, the best function value over the (N‑dimensional) space of all possible laryngoscope blade shapes was found using a simulated annealing optimization algorithm2.
RESULTS
The best achievable forward space is strongly dependant on the required
insertion range. Results of optimizing blade shapes given realistic anatomical
parameters and a maximum eyeline deviation of five degrees are shown in the
figure.
DISCUSSION
These results suggest that it is possible to design a better laryngoscope blade. Available blades look like the shapes returned by the optimization algorithm under the requirement of an insertion range of about twenty-five percent, which is consistent with the typical difference in size between blades of the same type. A blade constructed so as to achieve the effect of a blade modeled with zero insertion range is able to provide a larger forward space (into which to displace the tongue) without a corresponding increase in the eyeline deviation. The mathematical model used to obtain these results is quantitative, and is capable of calculating the effect of modifying the anatomical parameters. It is implemented in object-oriented C++ computer code, which is structured as a reusable framework for creating and testing other models, and for exploring different evaluation functions.
REFERENCES
2. Numerical Recipes in C: The Art of Scientific Computing. 2nd Edition
4. Br J Anaesth 1990; 65: 583-4P