Use of Preoperative Salivary Cortisol AM/PM Ratio for Prediction of Early Postoperative Cognitive Dysfunction

To the Editor In the recent article by Han et al1 assessing the predictive performance of preoperative salivary cortisol am/pm ratio for early postoperative cognitive dysfunction after noncardiac surgery in elderly patients, the receiver operating characteristic curve analysis showed that the area under the receiver operating characteristic curve for the salivary cortisol am/pm ratio in the postoperative cognitive dysfunction group was 0.72, and the optimal cutoff value was 5.69, with a sensitivity of 50% and a specificity of 91%. It should be pointed out that the area under the receiver operating characteristic curve is commonly interpreted as excellent (0.9–1), good (0.75–0.89), fair (0.6–0.74), low (0.5–0.59), or fail/no (<0.5) predictive ability.2 This indicates that preoperative salivary cortisol am/pm ratio can only provide a fair predictive ability for early postoperative cognitive dysfunction. Most important, readers were not provided with the Youden index at the optimal cutoff value and positive and negative predictive values of the preoperative salivary cortisol am/pm ratio for early postoperative cognitive dysfunction. Other than ability to provide the optimal cutoff value, the Youden index actually is a direct measure of diagnostic accuracy at the optimal cutoff value (ie, the maximum overall correct classification rate that a biomarker can achieve). Namely, even a biomarker with a large area under the receiver operating characteristic curve may have an unsatisfactory overall correct classification rate at the optimal cutoff point, and vice versa.2 Thus, by not providing the Youden index, positive and negative predictive values of the preoperative salivary cortisol am/pm ratio have resulted in difficulty determining whether it has a good discriminative power for early postoperative cognitive dysfunction. When the multivariate logistic regression analysis with forward likelihood ratio method was used to determine the association between the preoperative salivary cortisol am/pm ratio and the occurrence of early postoperative cognitive dysfunction, moreover, the postoperative adverse events and complications were not included in the model. In fact, postoperative infections, respiratory complications, and inadequate postoperative pain control have been identified as significant risk factors for early postoperative cognitive dysfunction after noncardiac surgery in elderly patients.3 We argue that not taking these postoperative risk factors into account would have distorted with the inferences of multivariate regression analyses for risk factors of early postoperative cognitive dysfunction and their adjusted odds ratios in this study. Fu Shan Xue, MDLiu Jia Zi Shao, MDRui Juan Guo, MDDepartment of AnesthesiologyBeijing Friendship HospitalCapital Medical UniversityBeijing, China[email protected]