Validierung von Score-Systemen zur Vorhersage einer schwierigen Maskenbeatmung - eine prospektive Studie

The significance of difficult or impossible mask ventilation becomes most evident when it occurs in cannot ventilate - cannot intubate situations. In two large studies conducted by Olivier Langeron in 2000 and by Sachin Kheterpal in 2006, the authors investigated risk factors for difficult mask ventilation (DMV) and recommended the use of score systems for preoperative risk evaluation. Receiver operating characteristic of the results revealed that, in the two studies, approximately 75% of the patients (76% in Langeron´s study and 75% in Kheterpal´s study) were correctly classified to the DMV category. Although the accuracy of prediction was not very high and no validation in an external population was performed in either study, the use of these score systems was recommended because of their simple handling. The aim of the present study was to investigate, whether the scores devised by Langeron and Kheterpal for assessment of mask ventilation can be usefully applied in clinical routine and the risk of anaesthesia for patients can be further reduced by this approach. Data pertaining to 3458 patients from various surgical specialities were included in the study. Difficult mask ventilation occurred in 7.3% of cases according to Langeron´s definition, and in 6.5% of cases according to Kheterpal´s criteria. We registered a high risk of DMV for the parameters of age, increased BMI, growth of a beard, impossible mandibular dislocation, snoring and edentulousness. However, as the parameters were poorly correlated with the actual event, none of them was suitable to predict DMV. Like the original studies, sum analysis also revealed that the frequency of difficult mask ventilation was markedly increased in direct proportion to the number of existing risk factors in a patient. The patient with the maximal number of risk factors had difficulties with mask ventilation ten times more frequently than did the patient with no risk factors. However, the ROC curve showed an area under the curve (AUC) of 0.64 for Langeron´s investigation, whereas the AUC according to Kheterpal´s definition was 0.60. Thus, the selectivity of these tests was found to be unacceptable. In both cases, the best prediction was observed at a cut-off of at least two risk factors in a patient. For this limit, sensitivity was 0.72 for Langeron´s data and only 0.62 for Kheterpal´s criteria. Furthermore, both studies were very unspecific (specificity, 0.5) and the significance of a positive test result low, because ventilation could eventually be performed without difficulties in more than 90% of the patients. Thus, determination of the risk factors determined by Langeron and Kheterpal and their use in the recommended risk scores gives the anaesthesist no reliable information about the occurrence of DMV. Therefore, the time required to determine the score is clearly not justified in clinical routine. At this moment we have to focus on intensive tutorial of each anaesthesist, as the absence of appropriate measuring methods increases the value of clinical experience.