We would like to thank Yamazaki and Hosokawa for their interest in our article.  The authors highlight four concerns about using the Swallowing, Tongue protrusion, Airway protection reflected by spontaneous and suctioning cough, and Glasgow coma scale Evaluation (STAGE) score to predict extubation readiness in neurosurgical patients.

Regarding the first concern, we agree that standardizing extubation criteria and definitions are expected for research consistency and comparability. However, heterogeneity in extubation protocols is common due to the lack of consensus. Therefore, it is not always feasible to require different centers to adopt identical extubation protocols or limit the postextubation support when some centers prefer to use it prophylactically. The time window used to define extubation failure ranges from 2 to 7 days.  While extending the time window to 5 to 7 days may capture more extubation failures, it is important to note that some reintubations are due to disease deterioration (e.g., postoperative intracranial bleeding), unrelated to the initial extubation assessment.

Second, we excluded direct tracheostomy since it was outside the scope of our study aim. Critically neurosurgical patients often present poor consciousness and weak airway protection function. We aimed to develop a model combining these two aspects to predict extubation readiness for patients after successful spontaneous breathing trials. Weaning difficulty can be easily identified during spontaneous breathing trials, which is beyond our aim. Low STAGE scores indicate a low likelihood of extubation success, promoting clinicians to determine whether to proceed with extubation at risk, prolonged ventilation, or direct tracheostomy.

Third, we agree that STAGE score items evaluation is associated with the sedation level and evaluation timing. We recommend evaluating each item daily at the same time (e.g., during early rounds) after sedation reduction or cessation, as this was also part of our extubation protocol. Patterns of airway dysfunction vary among different brain injury lesions and etiologies with different brain conduction circuits. We have observed that patients with infratentorial lesions often exhibit poor suctioning cough and/or swallowing but adequate spontaneous cough. In contrast, intracerebral hemorrhage and traumatic brain injury patients often have acceptable suctioning cough, and less saliva but poor spontaneous cough. Therefore, careful evaluation of each item is needed. In addition, extubation decisions were made by intensive care unit physicians and neurosurgeons based on their clinical experience without using the STAGE score. The STAGE score was assessed by two respiratory therapists without informing the clinicians to avoid interfering with extubation decisions. Quantitative measures of cough peak flow can enhance the credibility of the assessment but increase clinical complexity, which is more suitable for patients near the “extubation or not” threshold.

Fourth, as referenced by the traumatic brain injury phenotypes and its methodologywe believe that large-sample cohorts, detailed records, scoring systems development, and data integration across populations could help identify extubation failure phenotypes in neurosurgical patients. This could further facilitate standardizing extubation protocols and improving extubation decision-making.