© 2007 American Thoracic Society
The Diagnosis of Relative Adrenal Insufficiency: The Long and Winding Road...To the Editor:We have read with interest the article by Annane and colleagues regarding the definition of adrenal insufficiency (AI) (1), and we compliment the authors because they provided us with new and exciting data on the value of the cosyntropin test and levels of total and free cortisol. These data add important information to the discussion of the correct diagnostic tools for relative adrenal insufficiency in septic patients (2). However, we were puzzled by the results reported in Table 4 of Annane and colleagues' study (1). Apparently, we were not alone, and this has been addressed in a recent letter to the editor in the AJRCCM (3). In a reply, Annane and colleagues provided corrected values for the negative predictive value (NPV) of the different diagnostic cutoff levels for AI in Table 4 (4). However, we still have concerns regarding the conclusions reached in this article (1). Apparently, the NPV of baseline cortisol levels less than 10 µg/dl or delta cortisol less than 9 µg/dl was much lower than reported initially. There are therefore many false negatives, or many patients who actually have AI who are missed when this cutoff is used. We feel that the conclusion should be extended with a statement that a delta cortisol level greater than 9 µg/dl does not rule out adrenal insufficiency. This is an important issue as the delta cortisol cutoff of 9 µg/dl is currently the standard suggested in the Surviving Sepsis Campaign guidelines (5), and widely used in clinical practice. The second part of the conclusion, that AI is unlikely when the cosyntropin-stimulated cortisol level is 44 µg/dl or greater or delta cortisol is 16.8 µg/dl or greater, is intriguing. From the data provided in the online supplement (Table E3), we construct here Tables 1 and 2. The average positive predictive values in Table 2 question the reliability of using these cutoffs to exclude AI as suggested by the authors; a considerable number of patients did have values above the proposed cutoffs. The relatively poor performance of stimulated cortisol and delta cortisol levels can also be derived from the receiver operating characteristic (ROC) curves available in the online supplement (area under the ROC curves: 0.69 and 0.73, respectively) We feel that the results of this study should be carefully reinterpreted, and the conclusions reconsidered.
Ghent University Hospital, Ghent, Belgium FOOTNOTES Conflict of Interest Statement: Neither author has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. REFERENCES
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||