Commentary: Early Versus Delayed Extubation After Elective Neurosurgery for Brain Metastases

A routine change in when patients come off breathing machines after surgery for brain metastases may be less settled than it first appears. A new critical appraisal published in the Journal of Cancer Research and Clinical Oncology argues that evidence supporting early postoperative extubation is promising but not strong enough to show that extubation timing carries no independent risk. The authors say the key problem is not simply that the underlying study was observational. Rather, its comparison may be distorted by changes in neurosurgical and intensive-care practice that occurred over time, while important measures of patient vulnerability and blood loss were not clearly incorporated into the final statistical model.

The debate concerns patients undergoing elective surgery to remove brain metastases, cancers that have spread to the brain from another part of the body. During a craniotomy, patients are placed under general anesthesia and connected to a mechanical ventilator through an endotracheal tube. At the end of surgery, clinicians must decide whether to remove that tube promptly or continue ventilation in an intensive-care unit and extubate later. Early extubation can allow faster neurological assessment, reduce exposure to ventilator-associated complications and potentially shorten intensive-care treatment. Delayed extubation may be chosen when there is concern about swelling, bleeding, prolonged anesthesia, respiratory instability, substantial blood loss or an uncertain neurological examination. The decision is therefore both a respiratory intervention and a strategy for managing the immediate risks of brain surgery.

The commentary, by Xiaomei Wu of Pingliang Rehabilitation Center Hospital and Lihua Chen of Lanzhou Second People’s Hospital in China, examines a 2025 study by Khalafov and colleagues that compared early and delayed extubation after elective neurosurgical treatment of brain metastases. That earlier study reported no significant difference in adverse events between the two approaches and explored factors associated with extubation-related outcomes. Wu and Chen describe the work as clinically useful, particularly because dedicated evidence on extubation timing after brain-metastasis resection remains limited. But they caution that the findings should be viewed as hypothesis-generating rather than definitive evidence that early and delayed extubation are interchangeable for all patients.

Their first concern is the study’s use of a historical control group. The early-extubation protocol began in 2019, whereas the delayed-extubation comparison group consisted of patients treated during the preceding year under the hospital’s earlier routine intensive-care protocol. This is a before-and-after historical cohort, not a concurrent comparison in which patients treated during the same period receive different strategies. The distinction is crucial because the calendar year becomes entangled with the treatment. If the patients extubated early were treated after a change in protocol, any difference—or lack of difference—in complications could reflect the protocol, the passage of time, or both.

Medical care rarely stays constant while a study is under way. Between 2018 and 2020, clinicians may have changed anesthetic agents, reduced opioid use, refined postoperative pain control, altered nursing coverage or become more experienced with early mobilization and enhanced-recovery pathways. Even apparently modest improvements can influence outcomes after craniotomy. An opioid-sparing anesthetic, for example, may reduce sedation and respiratory depression, making prompt extubation easier. Better pain management may support coughing and breathing without excessive medication. More experienced staff may detect neurological deterioration earlier or mobilize patients sooner. These factors could make the later early-extubation group appear safer even if extubation timing itself had little effect.

The same issue applies to a finding of equivalent risk. If adverse-event rates were similar in the historical groups, researchers cannot cleanly determine whether early extubation truly produced outcomes comparable to delayed extubation or whether improvements in other aspects of care compensated for the change. In epidemiology, this is often described as confounding by secular trends: factors associated with time change alongside the exposure being studied. Wu and Chen suggest that reporting outcomes separately by surgical year, or limiting the analysis to a period in which patients from both strategies were treated concurrently, could help distinguish the effect of extubation timing from broader evolution in perioperative care. A prospective study would go further by defining eligibility and outcome measures before surgery and applying the extubation strategy under a controlled protocol.

The second major criticism involves the study’s multivariable logistic regression. Logistic regression estimates the association between one or more predictors and a binary outcome—such as whether a complication occurred—while statistically accounting for other variables in the model. Khalafov and colleagues included age, sex, American Society of Anesthesiologists physical-status score, tumor location and preoperative Karnofsky Performance Status. These variables capture demographic features, general health, disease location and functional capacity. However, Wu and Chen note that the Charlson Comorbidity Index was not included, even though the original authors observed numerical differences in comorbidity burden between the early- and delayed-extubation groups and discussed comorbidity as relevant to postoperative risk.

The Charlson Comorbidity Index combines chronic diseases into a weighted score intended to summarize a patient’s burden of illness. It is not a perfect measure of surgical risk, and it may overlap with other clinical variables, but excluding it can matter when comparing groups with different underlying health profiles. A patient with substantial cardiac, pulmonary, renal or metabolic disease may be more likely to require prolonged ventilation or intensive monitoring regardless of whether extubation is attempted early. If that burden is unevenly distributed between comparison groups and is not adjusted for, the apparent effect of extubation timing may be shifted in either direction. The problem is not that every possible variable must enter a model, but that the rationale for including or excluding clinically important covariates must be transparent.

Blood loss and transfusion create a related challenge. The commentary says intraoperative transfusion was the only variable to reach statistical significance in univariate analysis, which examines each predictor separately in relation to the outcome. Yet transfusion was not listed among the variables stated as inputs to the multivariable model, despite being described as tested. That leaves uncertainty about whether transfusion status was fully adjusted for, assessed later in a post hoc analysis or handled in another way. This distinction has practical consequences. Major bleeding can signal a more complex operation, physiological instability, coagulation problems or a greater need for postoperative support. Transfusion itself may also be associated with inflammation and complications, although an observational study cannot automatically separate the effect of the transfusion from the severity of the event that prompted it.

Despite these limitations, the commentary does not reject early extubation. Instead, it narrows the clinical conclusion. The available evidence appears most supportive of removing the breathing tube soon after surgery in carefully selected patients who have a stable intraoperative course, limited blood loss and no immediate signs of respiratory or neurological instability. Across the study under discussion and other craniotomy cohorts cited by the authors, blood loss and transfusion requirements seem more closely linked to delayed extubation decisions than tumor location or a comorbidity score alone. That pattern is biologically plausible: bleeding, prolonged surgery and physiological stress can increase the risk that a patient will not maintain adequate ventilation or will need continued airway protection.

The warning is against turning that provisional signal into a universal rule. Patients with greater comorbidity, significant transfusion requirements, difficult airway management, impaired consciousness or concern about cerebral swelling may not share the same risk profile as stable patients selected for early extubation. Extubation requires more than adequate oxygen levels at the end of anesthesia. Clinicians typically assess spontaneous breathing, carbon-dioxide clearance, airway reflexes, muscle strength, hemodynamic stability and the ability to protect the airway. After brain surgery, they must also consider whether sedation or neurological impairment could mask deterioration. A breathing tube can support ventilation, but it can also delay neurological examination and expose a patient to complications such as airway trauma, ventilator-associated pneumonia and weakness from prolonged critical illness. The decision is therefore a balance, not a contest between a universally safe and a universally dangerous option.

Wu and Chen conclude that a stronger evidence base will require a concurrently controlled or prospective design, with eligibility criteria, extubation protocols, complications and adjustment variables specified clearly. Future analyses should report the full set of candidate covariates, explain how missing data were handled and include comorbidity burden and transfusion status when clinically justified. Researchers should also distinguish factors that influence the decision to delay extubation from factors that actually cause poor outcomes, because the two are not identical. Until such studies are available, the 2025 findings offer reassurance about feasibility in selected patients, but they do not establish that extubation timing has no independent effect. In a field where a few hours of ventilation can influence both neurological monitoring and critical-care exposure, that distinction may determine whether an apparently simple protocol change is safely generalized.