| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
Department of Anesthesiology and Pain Medicine; naguib{at}mdanderson.org (Naguib) Department of Biostatistics and Applied Mathematics; University of Texas M. D.; Anderson Cancer Center; Houston, TX (Ensor) Department of Anesthesia; Roy J. and Lucille A. Carver College of Medicine; University of Iowa; Iowa City, IA (O’Sullivan)
In Response:
Dr. van Klei et al.’s letter (1) reflects a number of misunderstandings and misconceptions regarding our study (2), which needs clarification. The aim of our study was not to estimate or predict the absolute probability of difficult intubation, as they have stated. Our study tested the discriminative power of three difficult intubation models (3–5). Age, weight, height, and sex could affect intubation conditions; therefore, a matched case–control study was used to account for these confounding variables.
With respect to Dr. van Klei et al.’s first concern, a blinded investigator collected objective data, as described in our article, and the fact that he knew that one of the patients had a difficult airway should not have affected these objective measurements. We do not believe that the low incidence of difficult intubation reported in our study was due to under-reporting. In our article (2), we clearly discussed the reason for a low incidence of difficult intubation. We stated ".... this study may have increased practitioner awareness of difficult airways in patients presenting for surgery and prompted more aggressive use of alternate airway-management techniques, leading to a decrease in the incidence of unanticipated difficult intubation." We also noted these Hawthorne effects are a primary weakness of self-reporting studies.
We agree with Dr. van Klei et al. that positive predictive value (PPV) and negative predictive value (NPV) cannot be directly estimated from a matched case–control study design. For this reason, the PPV and NPV calculations that we reported were based on a prevalence of difficult intubation of 5.8%, as reported in a recent meta-analysis (6). Therefore, the calculations that Dr. van Klei et al. made regarding their predictor [Mallampati score (MPI)] are not germane to our study and would not affect our conclusions. All MPI calculations Dr. van Klei et al. made were based on PPV and NPV. In contrast, our conclusions were based on sensitivity and specificity that are not affected by choosing a match-case control design. Thus, our comparison of the three models’ receiver operating characteristic curves yields an unprejudiced analysis of the performance for these three classification models. As stated in our article, our new model "must be prospectively validated." We would welcome such a comparison of the three models on an independent patient population.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  M. Naguib The Ideal Dose of Succinylcholine for Tracheal Intubation Anesth. Analg., December 1, 2006; 103(6): 1582 - 1582. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
