Correspondence to Professor Jian-Xin Zhou; [email protected]
STRENGTHS AND LIMITATIONS OF THIS STUDY
This study explored the predictive value of cough peak flow rate (CPF) for extubation in critically ill postcraniotomy patients for the first time.
For conscious patients, we first compared the predictive value of voluntary and involuntary CPF for extubation in critically ill patients after craniotomy.
This study also compared the accuracy of CPF with conventional extubation prediction parameters.
This was a single-centre study and the results may not be directly generalised to other centres.
Introduction
Endotracheal extubation is in postcraniotomy critically ill patients might be challenging due to uncertainties about persistent central nervous system dysfunction postextubation. Premature extubation may lead to an increased risk of reintubation as well as the risk of airway damage and hospital-acquired pneumonia, resulting in prolonged hospital stay and even adverse effects on neurological prognosis and mortality.1 However, delaying extubation may be also associated with increased morbidity in the form of side effects from analgesics and sedatives, nosocomial infections, increased mortality and higher costs due to prolonged length of stays (LOS) in intensive care unit (ICU) and in hospital.2 3 There is a need for variables with robust predictive value for successful extubation in these patients.
Previous studies found that conventional predictors of successful weaning, such as rapid shallow breathing index (RSBI), negative inspiratory force, minute ventilation, might to inaccurate in predicting extubation outcomes of patients with acute brain injury.4 5 For patients with impaired consciousness and/or insufficient cough strength to guard against aspiration during spontaneous breathing, the need for ventilatory support and an artificial airway may be separate.
Clinicians often face the dilemma of whether to extubate a brain-injured patient who meets all extubation criteria but has impaired consciousness and compromised ability to maintain airway patency. Coplin et al found that the reintubation rate was similar in patients with different Glasgow Coma Scale (GCS) scores (10 vs 7), but pre-extubation spontaneous cough and suctioning frequency were significantly associated with extubation failure.3 Their findings suggested that a patient with impaired consciousness might be extubated successfully if the ability to maintain upper airway patency is sufficient which mainly includes cough reflex and swallowing function.
Coughing and swallowing have been shown to share neuroanatomical substrates, including efferent nerves6 7 and central pattern generators located in the brainstem.7–10 Clinical studies have also found that dystussia (impaired cough) and dysphagia often coexist in patients with neurological diseases.11–15 We infer that dystussia or atussia (absence of cough) could predict swallowing dysfunction as well as the failure of endotracheal extubation.
The strength of cough has been used to predict the extubation outcomes of critically ill patients.16–19 However, its accuracy in predicting extubation outcomes in patients after craniotomy has not been well investigated. The purpose of this study was to evaluate the predictive value of the cough peak flow (CPF) for extubation success in postcraniotomy critically ill patients.
Methods
Study design and population
This was a prospective cohort study conducted in three ICUs (71 beds) of a teaching hospital between August 2019 and January 2021. The study was registered at ClinicalTrials.gov (NCT04000997).
Postcraniotomy patients were screened consecutively for eligibility. Patients who were 18 years or older, stayed in ICU for more than 24 hours and underwent mechanical ventilation for more than 24 hours were eligible for the study. Patients were excluded if one of the following was present: no extubation attempt during the ICU stay; underwent tracheostomy without extubation attempt; pregnant or lactating women; enrolled in other clinical trials; declined to participate in the study.
Data collection
Data were collected by two physicians. Demographic data regarding age, sex, height, weight, primary diagnosis, chronic comorbidities and information about the surgery (operative time and surgical site) were recorded. The Acute Physiology and Chronic Health Evaluation (APACHE) II score,20 hospital mortality, duration of mechanical ventilation, ICU LOS, hospital LOS and Glasgow Outcome Scale (GOS) at hospital discharge were calculated. GCS (language score was 1 point by default for the patient with an endotracheal tube)3 and Sequential Organ Failure Assessment score21 before extubation were recorded. The levels of haemoglobin and albumin were measured on the morning of extubation. An arterial blood gas analysis was performed prior to extubation and the results, including pH, the arterial oxygen partial pressure (PaO2), the partial pressure of carbon dioxide (PaCO2) and the ratio of PaO2 in mmHg to fractional inspired oxygen (FiO2) (P/F ratio) were recorded. The fluid balance 24 hours prior to extubation was calculated.
Process of weaning from mechanical ventilation
Our centre has an established protocol for weaning and extubation.22 Spontaneous breathing trial (SBT) was performed using a T-piece in this study and supplemental oxygen was provided with less than 40% of FiO2 through the T-piece. A Venturi device was placed between the oxygen supply and the T-piece to regulate oxygen concentration and reduce dead space.
During the first minute of SBT, a heated Fleisch pneumotachograph (Vitalograph, Lenexa, Kansas, USA) was placed between the T-piece and the endotracheal tube to record the flow tracings continuously as well as the respiratory rate (RR). The airway pressures were also measured continuously by pressure transducers (KT 100D-2, KleisTEK, Italy, range: ±100 cmH2O) connected to an ICU-Lab Pressure Box (ICU Lab, KleisTEK Engineering, Bari, Italy) by 80 cm rigid tube lines. Flow and pressure signals were displayed continuously and saved (ICU-Lab 3.0 Software Package, ICU Lab, KleisTEK Engineering, Bari, Italy) in a laptop for further analysis, at a sample rate of 200 Hz. The tidal volume (Vt) was calculated based on the area under the flow-time curve. RSBI was the ratio of the RR (breaths per minute) to the average Vt (L) over 1 min.23 Then the pneumotachograph would be removed.
The SBT lasted for 30–120 min depending on the patient’s medical history and condition. If the SBT failed, mechanical ventilation was reinstituted. For patients passing SBT and with risk factors for postextubation stridor (including traumatic intubation, intubation more than 6 days, large endotracheal tube, female sex and reintubation after unplanned extubation), a cuff leak test would be performed and a cut-off of 110 mL was used to determine the result of test.24 25 Patients with failed cuff leak tests would receive systemic steroids at least 4 hours before extubation. The implementation of SBT, assessment of SBT results and decisions regarding extubation were all determined by the patient’s attending teams. The extubation was performed by the respiratory therapist.
Extubation failure was defined as the need for reintubation within 72 hours after extubation. Reintubation was determined by the physician in charge and would be performed if at least one of the following criteria were met: (1) FiO2>50% to maintain pulse oxygen saturation (SpO2)>90%; (2) RR>35 breaths/min with accessory respiratory muscle involvement; (3) respiratory or cardiac arrest; (4) inadequate cough to clear airway secretions; (5) PaCO2 greater than 50 mm Hg with pH lower than 7.35; (6) neurological deterioration; (7) haemodynamic instability; (8) severe agitation or other clinical signs of respiratory fatigue. Reasons for reintubation were recorded. Laryngeal oedema was confirmed by laryngoscopy. For the patient with multiple extubation attempts during the same hospitalisation, only the first attempt was evaluated.
CPF measurement
CPF was measured by the Fleisch pneumotachograph by two trained physicians before extubation. The accuracy of CPF measurement with the Fleisch pneumotachograph has been confirmed in previous studies.26 Similar to RSBI measurements, when measuring CPF, the pneumotachograph was placed between the T-piece and the endotracheal tube, continuously recording flow and airway pressure which were saved to the laptop for further analysis.
Before CPF measurement, the patients were positioned at 30–45° in bed and secretions were removed by suction. The involuntary CPF (CPF-invol) was measured in all enrolled patients using the method described by Duan et al.16 This process involved removing the T-piece and dripping 2 mL of normal saline into the endotracheal tube at the end-inspiratory point to induce coughing. Then, the T-piece and pneumotachograph were reconnected to the endotracheal tube as soon as possible and real-time changes in airway pressure and airflow velocity were continuously recorded. When the patient’s breathing became smooth and regular, the measurement would be repeated at least twice using the same method.
Regardless of the eye-opening response, if a patient’s motor response score on the GCS was 6, the patient was considered conscious and their voluntary CPF (CPF-vol) would be measured using the same instrument as that for CPF-invol. Before measurement, the patients were allowed to rest for 10 min. The investigators would explain the procedure for coughing on command to the patients and perform a live demonstration. Specifically, the patients were instructed to inhale as deeply as possible and to exhale as quickly as possible through the endotracheal tube and pneumotachograph.16 26 At least three cough trials were conducted, with intervals of 30–60 s between trials. The mean of the three best values of voluntary and involuntary CPF were recorded.
The measured CPF values were blinded to the patients’ attending physicians and the extubation outcomes of patients were blinded to the physicians conducting CPF measurement and analysis. The investigators of this study were not involved in any treatment decisions of the patients. Inconsistencies and uncertainties in the data were resolved through interviews with the physicians responsible for data collection and verified by two chief physicians. If a patient’s CPF was not measured before extubation, such as in the case of unplanned extubation, the patient would not be included in the analysis.
Outcomes
The primary outcome of this study was the predictive accuracy of CPF for extubation in critically ill patients after craniotomy. Secondary outcomes included the predictive accuracy of other parameters, such as RSBI, APACHE II score, GCS, fluid balance, haemoglobin levels and albumin levels for extubation outcomes as well as comparisons between their accuracy and that of CPF in predicting extubation.
Statistical analysis
Categorical variables were expressed as frequencies and percentages and were analysed using χ² test or Fisher’s exact test. Continuous variables were expressed as the mean (±SD) or median (IQR) and were analysed using Student’s t-test or Mann-Whitney U-test, as appropriate. The optimal cut-off point was determined by the Youden index. The receiver operating characteristic (ROC) curve was used to assess the value of CPF and other parameters in predicting extubation success. The area under the ROC curve values (AUC), which reflect the diagnostic accuracy and predictive ability for successful extubation, were calculated for each parameter.
The sample size was estimated using PASS V.15.0 software. According to the data from this centre over the past year, the assumed extubation failure rate was 15.6%. Based on previous studies, we assumed that the AUC of CPF for predicting successful extubation was 0.7 and the AUC of the null hypothesis was 0.60. To detect the difference in AUCs between the null and alternative hypotheses, at least 673 patients were needed to achieve a power (1-β) of 0.9 and a Type I error probability (α) of 0.05.
We performed an internal validation of the accuracy of CPF in predicting successful extubation using the repeated k-fold (k=10) cross-validation method. The entire data set was divided into 10 folds of roughly equal size. In the first fold, the first part was used for validation and the remaining nine parts for training. Similarly, in the second fold, the second part was used for validation. This process was continued for the rest of the folding. Then, the 10-fold cross-validation algorithm was repeated 200 times. The mean value of AUC was calculated as well as the coefficient of variation.
Statistical analyses were performed using MedCalc software V.20.305 for Windows and R V.4.3.2 (www.R-project.org). All comparisons were unpaired. Two-tailed p values<0.05 were considered statistically significant. UpSet plots were used to depict the distribution of reasons for reintubation in patients with failure extubation and were implemented using the TBtools-II software.27
Results
Baseline demographics
During the study period, 4879 patients were admitted into the ICU wards and 1037 were eligible for the study, among whom 252 were excluded and 785 were included in the study (figure 1). The median age was 49 years. Nearly half of them (49.4%) were male (table 1). Hypertension, diabetes and cancer were the most common comorbidities of these patients. The primary reason for craniotomy was intracranial tumours (n=605, 77.1%), followed by cerebrovascular diseases (n=98, 12.5%), trauma (n=49, 6.2%), and other causes (n=33, 4.2%). Among intracranial tumours, meningiomas were the most common (n=194), followed by gliomas (n=121) and schwannomas (n=86). Regarding cerebrovascular diseases, vascular malformations (n=54) were the most common, followed by aneurysmal subarachnoid haemorrhage (n=32). The lesions were mostly located in the infratentorial region (n=499, 63.6%) and the majority of patients underwent elective surgery (n=712, 90.7%). The median surgery duration was 5.4 (IQR 4.0–7.0) hours.
Table 1
Characteristics of the patients
| Variables | Total (n=785) | Extubation success (n=641) | Extubation failure (n=144) | P value |
| Age (years)* | 49 (36–57) | 48 (36–56) | 53 (43–63) | <0.001 |
| Male | 388 (49.4%) | 315 (49.1%) | 73 (50.7%) | 0.736 |
| Height (m)* | 1.66 (1.60–1.72) | 1.66 (1.60–1.72) | 1.65 (1.60–1.71) | 0.328 |
| Weight (Kg)* | 65 (58–75) | 65 (58–75) | 65 (58–74) | 0.399 |
| BMI | 23.9 (21.5–26.2) | 23.9 (21.5–26.2) | 23.7 (21.3–26.2) | 0.645 |
| Smoking | 106 (13.5%) | 83 (12.9%) | 23 (16.0%) | 0.337 |
| Alcoholism | 38 (4.8%) | 30 (4.7%) | 8 (5.6%) | 0.658 |
| Comorbidities | ||||
| Hypertension | 161 (20.5%) | 114 (17.8%) | 47 (32.6%) | <0.001 |
| Diabetes | 67 (8.5%) | 46 (7.2%) | 21 (14.6%) | 0.004 |
| Cancer | 37 (4.7%) | 28 (4.4%) | 9 (6.3%) | 0.336 |
| Cerebrovascular disease | 17 (2.2%) | 13 (2.0%) | 4 (2.8%) | 0.532 |
| Coronary heart disease | 15 (1.9%) | 11 (1.7%) | 4 (2.8%) | 0.496 |
| Chronic lung disease | 5 (0.6%) | 4 (0.6%) | 1 (0.7%) | 1.000 |
| APACHE II* | 15 (12–18) | 14 (12–17) | 17 (15–21) | <0.001 |
| ICU readmission | 53 (6.8%) | 33 (5.1%) | 20 (13.9%) | <0.001 |
| Total duration of mechanical ventilation (day)* | 4 (2–11) | 2 (2–7) | 13 (9–20) | <0.001 |
| Duration of mechanical ventilation before extubation (day)* | 3 (2–7) | 2 (2–6) | 9 (6–14) | <0.001 |
| ICU LOS (day)* | 6 (2–14) | 4 (2–9) | 19 (13–29) | <0.001 |
| Hospital LOS (day)* | 20 (15–29) | 19 (14–26) | 31 (25–42) | <0.001 |
| GOS* | 5 (4–5) | 5 (4–5) | 3 (3–4) | <0.001 |
| Death | 21 (2.7%) | 8 (1.2%) | 13 (9.0%) | <0.001 |
*Data were expressed as median (IQR).
APACHE, Acute Physiology and Chronic Health Evaluation; BMI, body mass index; GOS, Glasgow Outcome Scale; ICU, intensive care unit; LOS, length of stay.
Among the included patients, 641 were successfully extubated and 144 failed (18.3%). The most common reason for reintubation was difficulty in clearing airway secretions (n=81), followed by excessive work of breathing (n=52) and refractory hypoxaemia (n=43) (the distribution of reasons for reintubation in patients with failure extubation was shown in online supplemental figure S1). Among the successfully extubated patients, four received non-invasive positive pressure ventilation (NPPV) and six received high-flow nasal cannula (HFNC) therapy. Among the patients who failed extubation, two received NPPV and one received HFNC therapy.
Patients with advanced age, impaired consciousness, comorbidities of hypertension (29.2% vs 15.5%, p<0.001) and diabetes (31.3% vs 17.1%, p=0.004), supratentorial lesions (23.8% vs 15.2%, p=0.003) and patients undergoing emergency surgery (41.1% vs 16.0%, p<0.001) had higher extubation failure rates. Patients with trauma were more likely to experience extubation failure (46.9%) than other patients (25.6%).
Compared with patients successfully extubated, those who experienced extubation failure had higher APACHE II scores, longer duration of mechanical ventilation, longer ICU and hospital LOS, higher ICU readmission rates, lower GOS scores and higher in-hospital mortality rates.
The successful extubation group had significantly higher GCS (10 vs 9, p<0.001), haemoglobin levels (118 vs 98 g/L, p<0.001) and albumin levels (35.3 vs 31.3 g/L, p<0.001) compared with the failed extubation group (online supplemental table S1). The P/F ratio was similar between the groups (264.9 vs 275.7, p=0.740), while the positive fluid balance in the 24 hours prior to extubation was lower in the successful group (0 vs 210 mL, p<0.001). Similarly, among conscious patients, those who successfully extubated had higher haemoglobin (119 vs 102 g/L, p<0.001) and albumin levels (35.4 vs 31.1 g/L, p<0.001), with no differences in the P/F ratio (267.6 vs 259.7, p=0.406) and a lower positive fluid balance (−20 vs 400 mL, p<0.001) (online supplemental table S2).
Predictive value of CPF-invol for successful extubation
For all patients, the median CPF-invol was 88.8 (IQR 67.2–99.0) L/min. CPF-invol was significantly higher in males (92.1 vs 84 L/min, p<0.001), conscious patients (90.0 vs 78.3 L/min, p<0.001), patients undergoing elective surgery (89.4 vs 75.6 L/min, p<0.001) as well as patients younger than 49 years (91.2 vs 87.6 L/min, p<0.001) (online supplemental figure S2). The CPF-invol of patients who were successfully extubated was significantly higher than that of patients who failed (91.2 vs 54.6 L/min, p<0.001) and the same was true when patients were grouped by age, gender, level of consciousness, type of surgery and site of lesion (figure 2).
Enlarge this image.Figure 2. Comparison of CPF-invol in patients with success and failure extubation grouped by age, gender, level of consciousness, type of surgery and site of lesion. * p<0.05; *** p<0.001. CPF-invol, involuntary cough peak flow; ns, no significance.
As shown in figure 3a, the AUC of CPF-invol for the prediction of successful extubation was 0.810 (95% CI 0.766 to 0.854), with a cut-off value of 63.2 L/min, a sensitivity of 87.4% and a specificity of 66.7% (table 2). The mean AUC of CPF-invol obtained by the 10-fold cross-validation was 0.810 (±0.069) and the coefficient of variation was 8.5%.
Enlarge this image.Figure 3. (a) Comparison of areas under the receiver operating characteristic curve (AUCs) of CPF-invol, pre-extubation GCS and RSBI in predicting extubation success in all patients. (b) Comparison of AUCs of CPF-invol, CPF-vol and RSBI in predicting successful extubation in conscious patients. CPF-invol, involuntary cough peak flow; CPF-vol, voluntary cough peak flow; GCS, Glasgow Coma Scale; RSBI, rapid shallow breathing index.
The predictive value of different parameters for successful extubation in postcraniotomy patients
| Variables | Cut-off level | Sensitivity (%) | Specificity (%) | Positive predictive value (%) | Negative predictive value (%) |
| All patients | |||||
| CPF-invol (L/min) | >63.2 | 87.4 | 66.7 | 92.1 | 54.2 |
| Age (years) | ≤47 | 49.1 | 68.8 | 87.5 | 23.3 |
| Pre-extubation GCS | >9 | 87.7 | 51.4 | 88.9 | 48.4 |
| APACHE II | ≤15 | 65.2 | 63.9 | 88.9 | 29.2 |
| RSBI | <54.42 | 75.4 | 45.1 | 85.9 | 29.1 |
| Haemoglobin (g/L) | >106 | 68.9 | 69.7 | 91.0 | 33.6 |
| Albumin (g/L) | >31.4 | 79.1 | 52.4 | 87.9 | 36.5 |
| Fluid balance (mL)* | ≤70 | 56.3 | 66.0 | 88.0 | 25.3 |
| Conscious patients | |||||
| CPF-invol (L/min) | >63.2 | 87.5 | 72.6 | 96.9 | 42.4 |
| CPF-vol (L/min) | >68.2 | 77.9 | 69.9 | 95.3 | 28.5 |
| Age (years) | ≤47 | 50.9 | 65.8 | 92.1 | 14.5 |
| APACHE II | ≤14 | 57.3 | 63.0 | 92.5 | 15.7 |
| RSBI | <47.9 | 63.8 | 61.6 | 92.9 | 17.7 |
| Haemoglobin (g/L) | >109 | 67.7 | 69.4 | 94.6 | 21.4 |
| Albumin (g/L) | >32.2 | 75.1 | 62.0 | 94.0 | 24.2 |
| Fluid balance (mL)* | ≤183 | 67.8 | 71.2 | 94.9 | 21.8 |
*Referred to the fluid balance within 24 hours before extubation.
APACHE, Acute Physiology and Chronic Health Evaluation; CPF-invol, involuntary cough peak flow; CPF-vol, voluntary cough peak flow; GCS, Glasgow Coma Scale; RSBI, rapid shallow breathing index.
Predictive value of CPF for successful extubation in conscious patients
A total of 651 (82.9%) patients were conscious, among whom 73 (11.2%) failed extubation. Their median CPF-invol and CPF-vol were 90.0 (IQR 72.0–99.6) L/min and 92.4 (IQR 66.0–106.8) L/min, respectively. There was poor consistency between CPF-vol and CPF-invol with a Spearman correlation coefficient of 0.27 (95% CI 0.19 to 0.34) (online supplemental figure S3). Both the CPF-invol (91.2 vs 47.5 L/min, p<0.001) and CPF-vol (93.0 vs 55.8 L/min, p<0.001) were significantly higher in the successful extubation group than in the failed extubation group.
The AUC of CPF-invol for the prediction of successful extubation was 0.849 (95% CI 0.794 to 0.904) (figure 3b) with a cut-off value of 63.2 L/min, a sensitivity of 87.5% and a specificity of 72.6% (table 2). The mean AUC of CPF-invol obtained by the 10-fold cross-validation was 0.852 (±0.086) and the coefficient of variation was 10.1%. The AUC of CPF-vol for the prediction of successful extubation was 0.756 (95% CI 0.696 to 0.817) with a cut-off value of 68.2 L/min, a sensitivity of 77.9% and a specificity of 69.9%. The mean AUC of CPF-vol obtained by the 10-fold cross-validation was 0.757 (±0.093) and the coefficient of variation was 12.4%. The AUC of CPF-invol was significantly higher than that of CPF-vol (p=0.019).
Patients with both high CPF-vol and CPF-invol were less likely to fail extubation (online supplemental figure S4). Patients with a high CPF-vol and a low CPF-invol were more likely to fail extubation than those with a low CPF-vol and a high CPF-invol.
Predictive value of other parameters for successful extubation
We also analysed the predictive value of the age (table 2 and online supplemental figure S5), pre-extubation GCS (figure 3), APACHE II, RSBI, haemoglobin, albumin and fluid balance in the 24 hours prior to extubation for predicting successful extubation. For all patients, their AUCs were 0.601, 0.712, 0.698, 0.630, 0.712, 0.682 and 0.633, respectively, all of which were lower than the AUC of CPF-invol (p<0.05). For conscious patients, the AUCs for predicting extubation success using the age, APACHE II, RSBI, haemoglobin, albumin and fluid balance in the 24 hours prior to extubation were 0.577, 0.616, 0.669, 0.704, 0.709 and 0.706, respectively, and were all lower than the AUC using CPF-invol (p<0.05).
Discussion
A total of 785 patients after craniotomy were included in the study, among whom 641 were successfully extubated. The extubation failure rate was 18.3%. CPF performed well in predicting extubation outcomes, whether measured voluntarily (in conscious patients only) or involuntarily (in all patients).
Postcraniotomy patients usually succeed in the SBT, but their extubation per se remains challenging. In our work, the extubation failure rate of patients who had passed the SBT was 18.3% and the result was in accordance with the failure rate values of 10–22.6% reported by previous literature those included neurocritical care patients2 3 28–31 and was similar to that in the extubation in neurocritical care patients (ENIO) study.31
Previous studies have also attempted to use airway scores to predict extubation outcomes in patients with brain injury. For example, Ibrahim et al 32 used a semi-quantitative cough strength score as a predictor for extubation outcome in traumatic brain injury and the results showed that 81.3% of patients with a score of 5 were successfully extubated, while all patients with a score of 0 were reintubated. Although the study demonstrated the effect of cough strength on extubation, it was difficult to predict extubation outcomes based on this classification in patients with a score of 2–4. In the ENIO study,31 the investigators used the Least absolute shrinkage and selection operator (LASSO) model containing 20 variables to predict extubation success with an AUC of 0.71 (95% CI 0.61 to 0.81) in the validation cohort. Given the large number of variables retained in the model, it may hinder its applicability and generalisability. They also tried to use a simplified score containing seven variables to predict extubation success, but the AUC dropped to 0.65 (95% CI 0.53 to 0.76). Xu et al developed a prediction score named spontaneous and suctioning cough, and Glasgow Coma Scale Evaluation (STAGE) score for the evaluation of extubation readiness in neurosurgical patients with mechanical ventilation.22 The result showed that the AUC for the total STAGE score to predict extubation success was 0.72 (95% CI 0.64 to 0.79). Although available data suggested that airway scores might predict extubation outcomes, the accuracy of the scores depends on the clinician’s experience. Strict and frequent training is necessary before they can be used, otherwise, they may be less accurate in predicting extubation outcomes. CPF has been used in previous studies to predict extubation outcomes. It is an objective measurement index with minimal influence from subjective factors and may be simpler to use and easier to promote widely compared with airway scores.
The cut-off values of CPF in previous studies varied from 35 L/min to 62.4 L/min and the accuracy in predicting extubation outcomes was inconsistent across studies,17 33–35 suggesting that the CPF might differ among patients with different diseases. As these studies rarely included neurocritical patients, their results might not be generalised to patients with acute brain injury. The study by Kutchak et al 36 found that patients with neurological diseases who have been mechanically ventilated for more than 24 hours had a higher risk of extubation failure if the CPF was less than 80 L/min (RR 3.6, 95% CI 2.0 to 6.7). The AUC of reflex cough in their study was 0.81 (95% CI 0.73 to 0.89). In our study, a CPF-invol>63.2 L/min could distinguish successful extubation with a sensitivity and a specificity of 87.3% and 66.7%, respectively. The cut-off value for CPF-invol in our study was similar to that of most previous studies. However, it was lower than that in Kutchak et al’s study, as was the mean CPF (84.0 L/min vs 102.09 L/min). The discrepancy in study populations might be the most important reason for differences in the results of the two studies. In Kutchak et al’s study, the proportion of male patients was higher than in ours (71.1% vs 49.1%). The higher CPF in men may explain why the mean CPF in their study was greater than in ours. Furthermore, most of our patients were conscious at the time of extubation and their GCS at extubation was higher than that of patients in Kutchak et al’s study. Both of our studies found that the higher the GCS, the greater the likelihood of successful extubation. This might be the reason why our patients had a higher extubation success rate despite a lower CPF.
Early clinical studies have found that some patients with brainstem lesions had a weakened cough reflex but could still clear the airway by voluntary cough, suggesting that there might be differences in pathways of voluntary and involuntary cough which has been confirmed by functional MRI.37 Mazzone et al explored the blood oxygen level-dependent responses to cough and found that the brain regions activated by voluntary and involuntary coughing were not exactly the same.37 Duan et al compared the predictive accuracy of voluntary and involuntary CPF for reintubation in cooperative patients and found that CPF-vol was more accurate than CPF-invol.16 However, the results of this study showed that the AUC of CPF-invol was significantly higher than that of CPF-vol. Although the measurement methods for CPF in the two studies were similar, differences in the equipment used to measure CPF might lead to differences in results.26 However, we speculate that the contrasting findings might be primarily due to differences in the studied populations. More than half of the patients in the study of Duan et al were intubated because of chronic obstructive pulmonary disease exacerbation. Their involuntary cough may be impaired but would not be absent. All of the patients in our study underwent craniotomy because of intracranial lesions. When measuring CPF, we found that in some ones the voluntary cough was normal, while the involuntary cough was very weak or even absent and there was a low correlation between voluntary and involuntary cough strength which has also been demonstrated in healthy individuals.38 As we know, the involuntary cough is automatically generated by afferent activation and is controlled by the brainstem.39 The voluntary cough is a conscious act that requires spontaneous activation of respiratory muscles and has been shown to be regulated by the cerebral cortices. Once these patients fell asleep, the protective effect of voluntary cough would be compromised, potentially leading to aspiration or ineffective airway clearance. Probably for this reason, patients in this study who had a strong voluntary cough and a weakened or absent involuntary cough were still more likely to fail extubation. Therefore, extubation should be done with caution in these patients.
The result of this study showed that the lower the GCS, the greater the likelihood of extubation failure. For example, the failure rate of extubation was as high as 87.5% in patients with a GCS≤5, while only 11.2% in patients with a GCS≥10. The study of Coplin et al 3 also evaluated the effect of GCS on extubation outcomes in patients with acute brain injury, but their findings differed from those of this study. They found that the GCS was associated with extubation delay but not with extubation failure. Therefore, they believed that extubation should not be delayed when impaired neurological status was the only concern for prolonged intubation. However, in the ENIO study, a GCS motor score=6 was an independent predictor of successful extubation.31 A post hoc analysis of the ENIO study was performed to identify the risk factors associated with extubation failure in patients with the most impaired level of consciousness.40 The results showed that the extubation failure rate was 25.9% in patients with GCS motor score≤5 which was significantly higher than that in patients with retained consciousness (19.6%, p=0.02). This result further confirmed that GCS might affect the extubation outcomes of neurocritical patients. A meta-analysis41 including the study of Coplin et al also found that a low GCS score was a risk factor for extubation failure, suggesting that the discrepancy between our findings and that of Coplin et al might be due to their relatively small sample sizes and insufficient statistical power to detect differences between groups.
Several studies have investigated the utility of conventional weaning parameters in the prediction of extubation failure in neurocritical care patients, such as RSBI. The study of Ko et al 5 included 62 patients with diverse neuropathologies, including stroke and trauma and the extubation failure rate was 17.7% which was similar to that of our patients. They found that RSBI could not predict extubation failure. The study of Anderson et al 28 also found no significant difference in RSBI between the successful and failed extubation groups. In our study, the RSBI of the successful extubation group was lower than that of the failed group. However, the RSBI only exhibited moderate sensitivity and poor specificity for predicting extubation success, similar to findings from a meta-analysis involving 10 946 patients.42 Moreover, the cut-off values for RSBI varied widely among different populations,42 suggesting that patients in different ICUs may have variable respiratory drive and clinical indications for reintubation and/or mechanical ventilation. For example, patients with an acute brain injury might have abnormal respiratory patterns, especially those with brain stem injury.43 44 Patients may have normal or high Vt, slow RRs and therefore small RSBI values. However, those patients may be reintubated due to hypoventilation. Different measurement methods might also lead to discrepancies between studies.
Previous studies have shown that haemoglobin, albumin and fluid balance are associated with extubation outcomes.45 46 The ROC analysis in this study also indicated that these parameters had some predictive value for extubation outcomes. However, their AUCs were significantly lower than that of CPF-invol. There was no difference in CPF-invol between trauma and non-trauma patients, but the extubation failure rate was higher in trauma patients. This might be related to the more critical condition of trauma patients, as reflected by a higher proportion of emergency surgeries (77.6% vs 4.8%, p<0.001), higher APACHE scores (20 vs 15, p<0.001), lower GCS prior to extubation (8 vs 10, p<0.001) and lower haemoglobin levels (96 vs 115 g/L, p<0.001). The high extubation failure rate in patients undergoing supratentorial surgeries might also be due to similar factors, as these patients tended to be older (52 vs 46 years, p<0.001), had higher APACHE scores (17 vs 14, p<0.001), lower GCS prior to extubation (10 vs 10, p<0.001), lower haemoglobin (102 vs 120 g/L, p<0.001) and albumin levels (33.4 vs 35.3 g/L, p<0.001) and a higher proportion of trauma cases (17.1% vs 0, p<0.001) and emergency surgeries (24.8% vs 0.4%, p<0.001).
This study has several limitations. First, this was a single-centre study and the decision to extubate a patient was jointly made by the attending intensivist and neurosurgeon responsible for the patient’s care. Our centre has extensive experience in managing critically ill postcraniotomy patients which might limit the generalisability of the study’s results to other centres. Second, three-quarters of the patients underwent craniotomy for brain tumours, while the number of patients with trauma and cerebrovascular diseases was relatively small. Most patients had elective surgeries with only a few emergency cases included. Additionally, most patients had infratentorial lesions with fewer patients having supratentorial lesions. Due to the insufficient number of cases, subgroup analyses for different craniotomy indications, surgical sites and types of surgery could not be performed which might also affect the extrapolation of this study. Third, although we adopted the CPF-invol measurement method described in previous literature, its accuracy and reliability have not been fully validated. It remains unclear whether different methods or varying doses of saline used to induce CPF could lead to changes that impact the study results. Future research could further explore this issue. Fourth, although CPF measurement is quite simple and does not require extensive training for the measurer, it does require specific equipment which may limit its widespread adoption and application. Additionally, sputum volume may affect extubation outcomes, but this study did not collect sputum volume data. According to our centre’s extubation protocol, patients with large amounts of sputum, especially those requiring repeated bronchoscopic suctioning, would not be considered for extubation. Lastly, we did not compare the accuracy of CPF with airway scores in predicting extubation outcomes. Future research could be conducted to compare their respective accuracies.
Conclusions
The results of our study suggested that the CPF might be reliable for the prediction of extubation success in postcraniotomy patients. Age, neurological status, APACHE II, RSBI, haemoglobin, albumin and fluid balance should also be assessed to further increase the likelihood of successful extubation.
Data availability statement
Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information.
Ethics statements
Patient consent for publication
Not applicable.
Ethics approval
This study involves human participants and was approved by the institutional review boards of Beijing Tiantan Hospital, Capital Medical University (KY 2019-063-02). Participants gave informed consent to participate in the study before taking part.
JZ and H-LL contributed equally.
Contributors JZ, H-LL and J-XZ designed the study, conducted the statistical analysis, interpreted the results, drafted and critically revised the manuscript. J Z and X-YL contributed in data collection. X-YL, G-QC, Y-LY, LZ, MX and GZS contributed in data analysis, interpretation of data and drafting the manuscript. All authors read and approved the final manuscript. The first author (JZ) is the guarantor.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
1 Reis HFC dos, Almeida MLO, Silva MF da, et al. Extubation failure influences clinical and functional outcomes in patients with traumatic brain injury. J Bras Pneumol 2013; 39: 330–8. doi:10.1590/S1806-37132013000300010
2 McCredie VA, Ferguson ND, Pinto RL, et al. Airway Management Strategies for Brain-injured Patients Meeting Standard Criteria to Consider Extubation. A Prospective Cohort Study. Ann Am Thorac Soc 2017; 14: 85–93. doi:10.1513/AnnalsATS.201608-620OC
3 Coplin WM, Pierson DJ, Cooley KD, et al. Implications of extubation delay in brain-injured patients meeting standard weaning criteria. Am J Respir Crit Care Med 2000; 161: 1530–6. doi:10.1164/ajrccm.161.5.9905102
4 dos Reis HFC, Almeida MLO, da Silva MF, et al. Association between the rapid shallow breathing index and extubation success in patients with traumatic brain injury. Rev Bras Ter Intensiva 2013; 25: 212–7. doi:10.5935/0103-507X.20130037
5 Ko R, Ramos L, Chalela JA. Conventional weaning parameters do not predict extubation failure in neurocritical care patients. Neurocrit Care 2009; 10: 269–73. doi:10.1007/s12028-008-9181-9
6 Pitts T, Morris K, Lindsey B, et al. Co-ordination of cough and swallow in vivo and in silico. Exp Physiol 2012; 97: 469–73. doi:10.1113/expphysiol.2011.063362
7 Davenport PW, Bolser DC, Morris KF. Swallow remodeling of respiratory neural networks. Head & Neck 2011; 33: S8–13. doi:10.1002/hed.21845
8 Bolser DC, Davenport PW. Functional Organization of the Central Cough Generation Mechanism. Pulm Pharmacol Ther 2002; 15: 221–5. doi:10.1006/pupt.2002.0361
9 Gestreau C, Milano S, Bianchi AL, et al. Activity of dorsal respiratory group inspiratory neurons during laryngeal-induced fictive coughing and swallowing in decerebrate cats. Exp Brain Res 1996; 108: 247–56. doi:10.1007/BF00228098
10 Jean A. Brain stem control of swallowing: neuronal network and cellular mechanisms. Physiol Rev 2001; 81: 929–69. doi:10.1152/physrev.2001.81.2.929
11 Plowman EK, Watts SA, Robison R, et al. Voluntary Cough Airflow Differentiates Safe Versus Unsafe Swallowing in Amyotrophic Lateral Sclerosis. Dysphagia 2016; 31: 383–90. doi:10.1007/s00455-015-9687-1
12 Pitts T, Bolser D, Rosenbek J, et al. Voluntary cough production and swallow dysfunction in Parkinson’s disease. Dysphagia 2008; 23: 297–301. doi:10.1007/s00455-007-9144-x
13 Hegland KW, Okun MS, Troche MS. Sequential voluntary cough and aspiration or aspiration risk in Parkinson’s disease. Lung 2014; 192: 601–8. doi:10.1007/s00408-014-9584-7
14 Smith Hammond CA, Goldstein LB, Zajac DJ, et al. Assessment of aspiration risk in stroke patients with quantification of voluntary cough. Neurology (ECronicon) 2001; 56: 502–6. doi:10.1212/wnl.56.4.502
15 Pitts T, Troche M, Mann G, et al. Using voluntary cough to detect penetration and aspiration during oropharyngeal swallowing in patients with Parkinson disease. Chest 2010; 138: 1426–31. doi:10.1378/chest.10-0342
16 Duan J, Liu J, Xiao M, et al. Voluntary is better than involuntary cough peak flow for predicting re-intubation after scheduled extubation in cooperative subjects. Respir Care 2014; 59: 1643–51. doi:10.4187/respcare.03045
17 Beuret P, Roux C, Auclair A, et al. Interest of an objective evaluation of cough during weaning from mechanical ventilation. Intensive Care Med 2009; 35: 1090–3. doi:10.1007/s00134-009-1404-9
18 Salam A, Tilluckdharry L, Amoateng-Adjepong Y, et al. Neurologic status, cough, secretions and extubation outcomes. Intensive Care Med 2004; 30: 1334–9. doi:10.1007/s00134-004-2231-7
19 Thille AW, Boissier F, Ben Ghezala H, et al. Risk factors for and prediction by caregivers of extubation failure in ICU patients: a prospective study. Crit Care Med 2015; 43: 613–20. doi:10.1097/CCM.0000000000000748
20 Knaus WA, Draper EA, Wagner DP, et al. APACHE II: a severity of disease classification system. Crit Care Med 1985; 13: 818–29.
21 Vincent J-L, de Mendonca A, Cantraine F, et al. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units. Crit Care Med 1998; 26: 1793–800. doi:10.1097/00003246-199811000-00016
22 Xu S-S, Tian Y, Ma Y-J, et al. Development of a Prediction Score for Evaluation of Extubation Readiness in Neurosurgical Patients with Mechanical Ventilation. Anesthesiology 2023; 139: 614–27. doi:10.1097/ALN.0000000000004721
23 Yang KL, Tobin MJ. A prospective study of indexes predicting the outcome of trials of weaning from mechanical ventilation. N Engl J Med 1991; 324: 1445–50. doi:10.1056/NEJM199105233242101
24 Kuriyama A, Jackson JL, Kamei J. Performance of the cuff leak test in adults in predicting post-extubation airway complications: a systematic review and meta-analysis. Crit Care 2020; 24: 640. doi:10.1186/s13054-020-03358-8
25 Girard TD, Alhazzani W, Kress JP, et al. An Official American Thoracic Society/American College of Chest Physicians Clinical Practice Guideline: Liberation from Mechanical Ventilation in Critically Ill Adults. Rehabilitation Protocols, Ventilator Liberation Protocols, and Cuff Leak Tests. Am J Respir Crit Care Med 2017; 195: 120–33. doi:10.1164/rccm.201610-2075ST
26 Bongers T, O’Driscoll BR. Effects of equipment and technique on peak flow measurements. BMC Pulm Med 2006; 6: 14. doi:10.1186/1471-2466-6-14
27 Chen C, Wu Y, Li J, et al. TBtools-II: A “one for all, all for one” bioinformatics platform for biological big-data mining. Mol Plant 2023; 16: 1733–42. doi:10.1016/j.molp.2023.09.010
28 Anderson CD, Bartscher JF, Scripko PD, et al. Neurologic examination and extubation outcome in the neurocritical care unit. Neurocrit Care 2011; 15: 490–7. doi:10.1007/s12028-010-9369-7
29 Karanjia N, Nordquist D, Stevens R, et al. A clinical description of extubation failure in patients with primary brain injury. Neurocrit Care 2011; 15: 4–12. doi:10.1007/s12028-011-9528-5
30 Asehnoune K, Seguin P, Lasocki S, et al. Extubation Success Prediction in a Multicentric Cohort of Patients with Severe Brain Injury. Anesthesiology 2017; 127: 338–46. doi:10.1097/ALN.0000000000001725
31 Cinotti R, Mijangos JC, Pelosi P, et al. Extubation in neurocritical care patients: the ENIO international prospective study. Intensive Care Med 2022; 48: 1539–50. doi:10.1007/s00134-022-06825-8
32 Ibrahim AS, Aly MG, Abdel-Rahman KA, et al. Semi-quantitative Cough Strength Score as a Predictor for Extubation Outcome in Traumatic Brain Injury: A Prospective Observational Study. Neurocrit Care 2018; 29: 273–9. doi:10.1007/s12028-018-0539-3
33 Su W-L, Chen Y-H, Chen C-W, et al. Involuntary cough strength and extubation outcomes for patients in an ICU. Chest 2010; 137: 777–82. doi:10.1378/chest.07-2808
34 Smina M, Salam A, Khamiees M, et al. Cough peak flows and extubation outcomes. Chest 2003; 124: 262–8. doi:10.1378/chest.124.1.262
35 Duan J, Zhou L, Xiao M, et al. Semiquantitative cough strength score for predicting reintubation after planned extubation. Am J Crit Care 2015; 24: e86–90. doi:10.4037/ajcc2015172
36 Kutchak FM, Debesaitys AM, Rieder M de M, et al. Reflex cough PEF as a predictor of successful extubation in neurological patients. J Bras Pneumol 2015; 41: 358–64. doi:10.1590/S1806-37132015000004453
37 Mazzone SB, Cole LJ, Ando A, et al. Investigation of the neural control of cough and cough suppression in humans using functional brain imaging. J Neurosci 2011; 31: 2948–58. doi:10.1523/JNEUROSCI.4597-10.2011
38 Mills C, Jones R, Huckabee ML. Measuring voluntary and reflexive cough strength in healthy individuals. Respir Med 2017; 132: 95–101. doi:10.1016/j.rmed.2017.09.013
39 Troche MS, Brandimore AE, Godoy J, et al. A framework for understanding shared substrates of airway protection. J Appl Oral Sci 2014; 22: 251–60. doi:10.1590/1678-775720140132
40 de Courson H, Massart N, Asehnoune K, et al. Risk factors of extubation failure in neurocritical patients with the most impaired consciousness. Intensive Care Med 2023; 49: 1251–3. doi:10.1007/s00134-023-07189-3
41 Wang S, Zhang L, Huang K, et al. Predictors of Extubation Failure in Neurocritical Patients Identified by a Systematic Review and Meta-Analysis. PLoS ONE 2014; 9: e112198. doi:10.1371/journal.pone.0112198
42 Trivedi V, Chaudhuri D, Jinah R, et al. The Usefulness of the Rapid Shallow Breathing Index in Predicting Successful Extubation: A Systematic Review and Meta-analysis. Chest 2022; 161: 97–111. doi:10.1016/j.chest.2021.06.030
43 Balofsky A, George J, Papadakos P. Neuropulmonology. Handb Clin Neurol 2017; 140: 33–48. doi:10.1016/B978-0-444-63600-3.00003-9
44 Punj P, Nattanmai P, George P, et al. Abnormal Breathing Patterns Predict Extubation Failure in Neurocritically Ill Patients. Case Rep Crit Care 2017; 2017: 9109054. doi:10.1155/2017/9109054
45 Li W, Zhang Y, Wang Z, et al. The risk factors of reintubation in intensive care unit patients on mechanical ventilation: A systematic review and meta-analysis. Intensive Crit Care Nurs 2023; 74: 103340. doi:10.1016/j.iccn.2022.103340
46 Li T, Zhou D, Zhao D, et al. Association between fluid intake and extubation failure in intensive care unit patients with negative fluid balance: a retrospective observational study. BMC Anesthesiol 2022; 22: 170. doi:10.1186/s12871-022-01708-3
© 2025 Author(s) (or their employer(s)) 2025. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ Group. http://creativecommons.org/licenses/by-nc/4.0/ This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/ . Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Details
; Hong-Liang, Li 2 ; Xu-Ying, Luo 1 ; Chen, Guang-Qiang 1 ; Yan-Lin, Yang 1 ; Zhang, Linlin 1 ; Xu, Ming 1 ; Shi, Guang Zhi 1 ; Zhou, Jian-Xin 3
1 Department of Critical Care Medicine, Beijing Tiantan Hospital, Capital Medical University, Beijing, Beijing, China
2 Department of Critical Care Medicine, Beijing Shijitan Hospital, Capital Medical University, Beijing, Beijing, China
3 Clinical and Research Center on Acute Lung Injury, Beijing Shijitan Hospital Capital Medical University, Beijing, Beijing, China