To the Editor:
Our thanks are due to Kheterpal et al.1 for another valuable addition to our understanding of airway management during anesthesia. Their work suggests that provided we exclude the patients with obvious abnormalities, we will find that the incidence of the combination of difficult mask ventilation (DMV) and difficult laryngoscopy (DL) is infrequent but not rare (0.4%), and the outcome is good when standard methods are applied.
As well as incidence and outcome, Dr. Kheterpal et al. addressed the prediction of DMV plus DL and presented odds ratios* to describe their findings. They used a group (class I, 0 to 3 risk factors) with few risk factors as a reference, but we should note that cases of DMV plus DL occurred (107 patients) in this group. The odds ratio for the group of patients with the most risk factors (class V, 7 to 11) was 18.4, which sounds high, but I suspect that the positive predictive value gives most of us a clearer idea of the clinical significance. The positive predictive value for DMV plus DL of class V abnormalities was only 3.31%, which means that 96.69% of predictions were wrong and the patients did not present difficulty. The conclusion must be that practitioners must expect to encounter such cases in low-risk patients and cannot be expected to predict the difficult cases with any accuracy.
What practitioners should be expected to do is to manage combined DMV and DL when it arises, so I am glad to see that this study tends to support the view that successful oxygenation, ventilation, and intubation are facilitated when the laryngeal “sphincter” is relaxed by neuromuscular blockade. Richardson and Litman2 have mentioned a “traditional anesthetic induction sequence taught on day 1 of residency,” which advises anesthesiologists to check that face-mask ventilation is possible before giving a relaxant drug. I believe that Kheterpal et al.’s report adds to the evidence pointing to the illogicality of this advice, which is actually of fairly recent and obscure origin.3
Competing Interests
The author declares no competing interests.
It is hard to know how best to combine the significance of associated conditions, which are usually described with odds ratios, and test results, which are often described with likelihood ratios (the ratio of true positives to false positives). For interest’s sake I considered the combination of items in class V as a test, calculated the likelihood ratio, and got a result of 8.9 (a value above 10 indicates clinical usefulness), which just restates the authors’ findings and confirms that practitioners cannot be expected to predict the problem cases with accuracy.
