Machine learning-based predictor for neurologic outcomes in patients undergoing extracorporeal cardiopulmonary resuscitation.

Background: We investigated the predictors of poor neurological outcomes in extracorporeal cardiopulmonary resuscitation (ECPR) patients using machine learning (ML) approaches. Methods: This study was a retrospective, single-center, observational study that included adult patients who underwent ECPR while hospitalized between January 2010 and December 2020. The primary outcome was neurologic status at hospital discharge as assessed by the Cerebral Performance Categories (CPC) score (scores range from 1 to 5). We trained and tested eight ML algorithms for a binary classification task involving the neurological outcomes of survivors after ECPR. Results: During the study period, 330 patients were finally enrolled in this analysis; 143 (43.3%) had favorable neurological outcomes (CPC score 1 and 2) but 187 (56.7%) did not. From the eight ML algorithms initially considered, we refined our analysis to focus on the three algorithms, eXtreme Gradient Boosting, random forest, and Stochastic Gradient Boosting, that exhibited the highest accuracy. eXtreme Gradient Boosting models exhibited the highest accuracy among all the machine learning algorithms (accuracy: 0.739, area under the curve: 0.837, Kappa: 0.450, sensitivity: 0.700, specificity: 0.740). Across all three ML models, mean blood pressure emerged as the most influential variable, followed by initial serum lactate, and arrest to extracorporeal membrane oxygenation (ECMO) pump-on-time as important predictors in machine learning models for poor neurological outcomes following successful ECPR. Conclusions: In conclusion, machine learning methods showcased outstanding predictive accuracy for poor neurological outcomes in patients who underwent ECPR.

Prognostic significance of respiratory quotient in patients undergoing extracorporeal cardiopulmonary resuscitation in Korea.

BACKGROUND: Respiratory quotient (RQ) may be used as a tissue hypoxia marker in various clinical settings but its prognostic significance in patients undergoing extracorporeal cardiopulmonary resuscitation (ECPR) is not known. METHODS: Medical records of adult patients admitted to the intensive care units after ECPR in whom RQ could be calculated from May 2004 to April 2020 were retrospectively reviewed. Patients were divided into good neurologic outcome and poor neurologic outcome groups. Prognostic significance of RQ was compared to other clinical characteristics and markers of tissue hypoxia. RESULTS: During the study period, 155 patients were eligible for analysis. Of them, 90 (58.1%) had a poor neurologic outcome. The group with poor neurologic outcome had a higher incidence of out-of-hospital cardiac arrest (25.6% vs. 9.2%, P=0.010) and longer cardiopulmonary resuscitation to pump-on time (33.0 vs. 25.2 minutes, P=0.001) than the group with good neurologic outcome. For tissue hypoxia markers, the group with poor neurologic outcome had higher RQ (2.2 vs. 1.7, P=0.021) and lactate levels (8.2 vs. 5.4 mmol/L, P=0.004) than the group with good neurologic outcome. On multivariable analysis, age, cardiopulmonary resuscitation to pump-on time, and lactate levels above 7.1 mmol/L were significant predictors for a poor neurologic outcome but not RQ. CONCLUSIONS: In patients who received ECPR, RQ was not independently associated with poor neurologic outcome.

Impact of early nutrition and feeding route on clinical outcomes of neurocritically ill patients.

Early proper nutritional support is important to critically ill patients. Nutritional support is also associated with clinical outcomes of neurocritically ill patients. We investigate whether early nutrition is associated with clinical outcomes in neurocritically ill patients. This was a retrospective, single-center, observational study including neurosurgical patients who were admitted to the intensive care unit (ICU) from January 2013 to December 2019. Patients who started enteral nutrition or parenteral nutrition within 72 hours after ICU admission were defined as the early nutrition group. The primary endpoint was in-hospital mortality. The secondary endpoint was an infectious complication. Propensity score matching (PSM) and propensity score weighting overlap weights (PSOW) were used to control selection bias and confounding factors. Among 1,353 patients, early nutrition was performed in 384 (28.4%) patients: 152 (11.2%) early enteral nutrition (EEN) and 232 (17.1%) early parenteral nutrition (EPN). In the overall study population, the rate of in-hospital mortality was higher in patients with late nutrition than in those with early nutrition (P<0.001). However, there was no significant difference in in-hospital mortality and infectious complications incidence between the late and the early nutrition groups in the PSM and PSOW adjusted population (all P>0.05). In the overall study population, EEN patients had a low rate of in-hospital mortality and infectious complications compared with those with EPN and late nutrition (P<0.001 and P = 0.001, respectively). In the multivariable analysis of the overall, PSM adjusted, and PSOW adjusted population, there was no significant association between early nutrition and in-hospital mortality and infectious complications (all P>0.05), but EEN was significantly associated with in-hospital mortality and infectious complications (all P<0.05). Eventually, early enteral nutrition may reduce the risk of in-hospital mortality and infectious complications in neurocritically ill patients.

Association between Fibrinogen-to-Albumin Ratio and Prognosis in Patients Admitted to an Intensive Care Unit.

The objective of this study was to investigate the usefulness of fibrinogen-to-albumin ratio (FAR) as a prognostic marker in patients admitted to an intensive care unit (ICU) compared with Sequential Organ Failure Assessment (SOFA) score, a widely used prognostic scoring system. An inverse probability weighting (IPW) was used to control for selection bias and confounding factors. After IPW adjustment, the high FAR group showed significantly higher risk of 1-year compared with low FAR group (36.4% vs. 12.4%, adjust hazard ratio = 1.72; 95% confidence interval (CI): 1.59-1.86; p < 0.001). In the receiver-operating characteristic curve analysis associated with the prediction of 1-year mortality, there was no significant difference between the area under the curve of FAR on ICU admission (C-statistic: 0.684, 95% CI: 0.673-0.694) and that of SOFA score on ICU admission (C-statistic: 0.679, 95% CI: 0.669-0.688) (p = 0.532). In this study, FAR and SOFA score at ICU admission were associated with 1-year mortality in patients admitted to an ICU. Especially, FAR was easier to obtain in critically ill patients than SOFA score. Therefore, FAR is feasible and might help predict long-term mortality in these patients.

Clinical Outcomes Associated with Degree of Hypernatremia in Neurocritically Ill Patients.

OBJECTIVE: Hypernatremia is a common complication encountered during the treatment of neurocritically ill patients. However, it is unclear whether clinical outcomes correlate with the severity of hypernatremia in such patients. Therefore, we investigated the impact of hypernatremia on mortality of these patients, depending on the degree of hypernatremia. METHODS: Among neurosurgical patients admitted to the intensive care unit (ICU) in a tertiary hospital from January 2013 to December 2019, patients who were hospitalized in the ICU for more than 5 days and whose serum sodium levels were obtained during ICU admission were included. Hypernatremia was defined as the highest serum sodium level exceeding 150 mEq/L observed. We classified the patients into four subgroups according to the severity of hypernatremia and performed propensity score matching analysis. RESULTS: Among 1146 patients, 353 patients (30.8%) showed hypernatremia. Based on propensity score matching, 290 pairs were included in the analysis. The hypernatremia group had higher rates of in-hospital mortality and 28-day mortality in both overall and matched population (both p<0.001 and p=0.001, respectively). In multivariable analysis of propensity score-matched population, moderate and severe hypernatremia were significantly associated with in-hospital mortality (adjusted odds ratio [OR], 4.58; 95% confidence interval [CI], 2.15-9.75 and adjusted OR, 6.93; 95% CI, 3.46-13.90, respectively) and 28-day mortality (adjusted OR, 3.51; 95% CI, 1.54-7.98 and adjusted OR, 10.60; 95% CI, 5.10-21.90, respectively) compared with the absence of hypernatremia. However, clinical outcomes, including in-hospital mortality and 28-day mortality, were not significantly different between the group without hypernatremia and the group with mild hypernatremia (p=0.720 and p=0.690, respectively). The mortality rates of patients with moderate and severe hypernatremia were significantly higher in both overall and matched population. Interestingly, the mild hypernatremia group of matched population showed the best survival rate. CONCLUSION: Moderate and severe hypernatremia were associated with poor clinical outcomes in neurocritically ill patients. However, the prognosis of patients with mild hypernatremia was similar with that of patients without hypernatremia. Therefore, mild hypernatremia may be allowed during treatment of intracranial hypertension using hyperosmolar therapy.

Prognostic Value of C-Reactive Protein and Albumin in Neurocritically Ill Patients with Acute Stroke.

We evaluated the prognostic value of C-reactive protein (CRP), albumin, CRP clearance (CRPc) and CRP/albumin ratio (CAR) in neurocritically ill patients with acute stroke. This is a retrospective, observational study. We included acute stroke patients who were hospitalized in the neurosurgical ICU from January 2013 to September 2019. The primary outcome was in-hospital mortality. A total of 307 patients were enrolled in the study. Among them, 267 (87.0%) survived until discharge from the hospital. CRP and CAR were significantly higher in non-survivors than in survivors (both p < 0.001). Serum albumin levels were significantly lower in the non-survivors than in the survivors (p < 0.001). In receiver operating characteristic curve analysis for prediction of in-hospital mortality, the area under the curve of CRP (C-statistic: 0.820) and CAR (C-statistic: 0.824) were greater than that of CRPc (C-statistic: 0.650) and albumin (C-statistic: 0.734) (all p < 0.005). However, there was no significant difference in the predictive performance between CRP and CAR (p = 0.287). In this study, CRP and CAR were more important than CRPc and albumin in predicting mortality of neurocritically ill patients with stroke. Early CRP level and CAR determination may help to predict the in-hospital mortality of these patients.

Association between High-Sensitivity Cardiac Troponin I and Clinical Prognosis of Neurosurgical and Neurocritically Ill Patients.

To investigate whether high-sensitivity troponin I (hs-TnI) elevation is associated with in-hospital mortality and major adverse cardiac events (MACEs) in neurosurgical and neurocritically ill patients. Among neurosurgical patients admitted to the intensive care unit (ICU) from January 2013 to December 2019, those whose serum hs-TnI levels were obtained within 7 days after ICU admission were included. Propensity score matching was used. Each patient with hs-TnI elevation was matched to a control patient. The primary endpoint was in-hospital mortality and the secondary outcome was MACEs. The hs-TnI elevation was shown in 848 (14.1%) of 6004 patients. After propensity score matching, 706 pairs of data were generated by 1:1 individual matching without replacement. In multivariable analysis of overall and propensity score-matched population, hs-TnI elevation was associated with in-hospital mortality (adjusted odds ratio (OR): 2.37, 95% confidence interval (CI): 1.68-3.33 and adjusted OR: 1.89, 95% CI: 1.28-2.81, respectively). In addition, hs-TnI elevation was associated with MACEs (adjusted OR: 2.73, 95% CI: 1.74-4.29 and adjusted OR: 2.64, 95% CI: 1.60-4.51, respectively). In this study, hs-TnI elevation was associated with in-hospital mortality and MACEs in neurosurgical and neurocritically ill patients.

Optimal Arterial Blood Gas Tensions for the Prognosis of Favorable Neurological Outcomes in Survivors after Extracorporeal Cardiopulmonary Resuscitation.

Our aim is to assess the optimal levels of oxygen and carbon dioxide for the prognosis of favorable neurologic outcomes in survivors after extracorporeal cardiopulmonary resuscitation (ECPR). We obtained the mean levels of PaCO2 and PaO2 in arterial blood gas samples 72 h after ECPR. The primary outcome was the neurological status, according to the Cerebral Performance Categories (CPC) scale, upon discharge. Of 119 (48.6%) survivors, 95 (38.8%) had favorable neurologic outcomes (CPC 1 or 2). There was a U-shaped relationship between mean arterial blood gas tensions and poor neurological outcomes. The risk of poor neurological outcome was lowest in patients with the second tertile of mean PaCO2 (30-42 mm Hg) and PaO2 (120-160 mm Hg). In a multivariable analysis, third tertile of mean PaCO2, third tertile of mean PaO2, age, shockable rhythm, out of hospital cardiac arrest, duration of cardiopulmonary resuscitation, and ECPR at cardiac catheterization lab were found to be significantly associated with poor neurologic outcomes. Additionally, hypercapnia and extreme hyperoxia were found to be significantly associated with poor neurological outcomes after ECPR. Therefore, maintaining adequate arterial levels of oxygen and carbon dioxide may be important for favorable neurological prognoses in survivors after ECPR.

Optimal Mean Arterial Pressure for Favorable Neurological Outcomes in Survivors after Extracorporeal Cardiopulmonary Resuscitation.

We evaluated the optimal mean arterial pressure (MAP) for favorable neurological outcomes in patients who underwent extracorporeal cardiopulmonary resuscitation (ECPR). Adult patients who underwent ECPR were included. The average MAP was obtained during 6, 12, 24, 48, 72, and 96 h after cardiac arrest, respectively. Primary outcome was neurological status upon discharge, as assessed by the Cerebral Performance Categories (CPC) scale (range from 1 to 5). Overall, patients with favorable neurological outcomes (CPC 1 or 2) tended to have a higher average MAP than those with poor neurological outcomes. Six models were established based on ensemble algorithms for machine learning, multiple logistic regression and observation times. Patients with average MAP around 75 mmHg had the least probability of poor neurologic outcomes in all the models. However, those with average MAPs below 60 mmHg had a high probability of poor neurological outcomes. In addition, based on an increase in the average MAP, the risk of poor neurological outcomes tended to increase in patients with an average MAP above 75 mmHg. In this study, average MAPs were associated with neurological outcomes in patients who underwent ECPR. Especially, maintaining the survivor's MAP at about 75 mmHg may be important for neurological recovery after ECPR.

Association between Neurologic Outcomes and Changes of Muscle Mass Measured by Brain Computed Tomography in Neurocritically Ill Patients.

This study aimed to investigate whether skeletal muscle mass estimated via brain computed tomography (CT) could predict neurological outcomes in neurocritically ill patients. This is a retrospective, single-center study. Adult patients admitted to the neurosurgical intensive care unit (ICU) from January 2010 to September 2019 were eligible. Cross-sectional areas of paravertebral muscles at the first cervical vertebra level (C1-CSA) and temporalis muscle thickness (TMT) on brain CT were measured to evaluate skeletal muscle mass. The primary outcome was the Glasgow Outcome Scale score at 3 months. Among 189 patients, 81 (42.9%) patients had favorable neurologic outcomes. Initial and follow-up TMT values were higher in patients with favorable neurologic outcomes compared to those with poor outcomes (p = 0.003 and p = 0.001, respectively). The initial C1-CSA/body surface area was greater in patients with poor neurological outcomes than in those with favorable outcomes (p = 0.029). In multivariable analysis, changes of C1-CSA and TMT were significantly associated with poor neurological outcomes. The risk of poor neurologic outcome was especially proportional to changes of C1-CSA and TMT. The follow-up skeletal muscle mass measured via brain CT at the first week from ICU admission may help predict poor neurological outcomes in neurocritically ill patients.

Severe pain-related adverse events of percutaneous dilatational tracheostomy performed by a neurointensivist compared with conventional surgical tracheostomy in neurocritically ill patients.

BACKGROUND: We evaluated severe pain-related adverse events (SAE) during the percutaneous dilatational tracheostomy (PDT) procedure performed by a neurointensivist and compared the outcomes with that of conventional surgical tracheostomy in neurocritically ill patients. METHODS: This was a retrospective and observational study of adult patients who were admitted to the neurosurgical intensive care unit between January 2014 and March 2018 and underwent tracheostomy. In this study, primary endpoints were incidence of SAE: cardiac arrest, arrhythmias, hypertension, hypotension, desaturation, bradypnea, or ventilatory distress. The secondary endpoint was procedure-induced complications. RESULTS: A total of 156 patients underwent tracheostomy during the study. Elective surgery of brain tumors (34.0%) and intracranial hemorrhage (20.5%) were the most common reasons for admission. The most common reasons for tracheostomy were difficult ventilator weaning or prolonged intubation (42.9%) and sedative reduction (23.7%). Tachycardia (30.1%) and hypertension (30.1%) were the most common SAE. Incidence of SAE was more common in conventional tracheostomy compared to PDT (67.1% vs. 42.3%, P = 0.002). The total duration of SAE (19.8 ± 23.0 min vs. 3.4 ± 5.3 min, P < 0.001) and procedural time (42.2 ± 21.8 min vs. 17.7 ± 9.2 min, P < 0.001) were longer in conventional tracheostomy compared to PDT. Multivariable adjustment revealed that only PDT by a neurointensivist significantly reduced the incidence of SAE by one third (adjusted odds ratio [OR]: 0.36, 95% confidence interval [CI]: 0.187-0.691). In addition, PDT by a neurointensivist deceased the duration of SAE by 8.64 min (ß: -8.64, 95% CI: - 15.070 - -2.205, P = 0.009) and prolonging the procedure time by every one minute significantly increased the duration of SAE by 6.38 min (ß: 6.38, 95% CI: 0.166-0.470, P < 0.001). Procedure-induced complications were more common in conventional tracheostomy compared to PDT (23.5% vs. 11.3%, P = 0.047). CONCLUSIONS: This retrospective and exploratory study of our single-center limited cohort of tracheostomy patients revealed that decreased SAE may be associated with short procedural time during the PDT procedure performed by a neurointensivist. It is proposed that PDT by a neurointensivist may be safe and feasible in neurocritically ill patients.

Prognostic Value of Early Intermittent Electroencephalography in Patients after Extracorporeal Cardiopulmonary Resuscitation.

The aim of this study was to investigate whether early intermittent electroencephalography (EEG) could be used to predict neurological prognosis of patients who underwent extracorporeal cardiopulmonary resuscitation (ECPR). This was a retrospective and observational study of adult patients who were evaluated by EEG scan within 96 h after ECPR. The primary endpoint was neurological status upon discharge from the hospital assessed with a Cerebral Performance Categories (CPC) scale. Among 69 adult cardiac arrest patients who underwent ECPR, 17 (24.6%) patients had favorable neurological outcomes (CPC score of 1 or 2). Malignant EEG patterns were more common in patients with poor neurological outcomes (CPC score of 3, 4 or 5) than in patients with favorable neurological outcomes (73.1% vs. 5.9%, p < 0.001). All patients with highly malignant EEG patterns (43.5%) had poor neurological outcomes. In multivariable analysis, malignant EEG patterns and duration of cardiopulmonary resuscitation were significantly associated with poor neurological outcomes. In this study, malignant EEG patterns within 96 h after cardiac arrest were significantly associated with poor neurological outcomes. Therefore, an early intermittent EEG scan could be helpful for predicting neurological prognosis of post-cardiac arrest patients after ECPR.

Clinical usefulness of early serial measurements of C-reactive protein as outcome predictors in patients with subarachnoid hemorrhage.

BACKGROUND: The purpose of this study was to evaluate the role of C-reactive protein (CRP) in predicting neurological outcomes of patients with subarachnoid hemorrhage (SAH). METHODS: In this retrospective, observational study of adult patients with SAH treated between January 2012 and June 2017. Initial CRP levels collected within 24 h from the onset of SAH, the follow-up CRP levels were measured. The primary outcome was neurological status at six-month follow-up assessed with the Glasgow Outcome Scale (GOS, 1 to 5). RESULTS: Among 156 patients with SAH, 145 (92.9%) survived until discharge. Of these survivors, 109 (69.9%) manifested favorable neurological outcomes (GOS of 4 or 5). Initial CRP levels on admission and maximal CRP levels within four days were significantly higher in the group with poor neurological outcome compared with those manifesting favorable neurological outcomes (P = 0.022, P < 0.001, respectively). However, the clearance of CRPs did not differ significantly between the two groups (P = 0.785). Analysis of the receiver operating characteristic curve for prediction of poor neurological outcome showed that the performance of the maximal CRP was significantly better compared with the initial CRP or the clearance of CRP (P = 0.007, P < 0.001, respectively). In this study, the effect of CRP on neurological outcomes differed according to surgical clipping. The maximal CRP levels within four days facilitate the prediction of neurological outcomes of SAH patients without surgical clipping (C-statistic: 0.856, 95% confidence interval [CI]: 0.767-0.921). However, they were poorly associated with neurological prognoses in SAH patients who underwent surgical clipping (C-statistic: 0.562, 95% CI: 0.399-0.716). Multivariable logistic regression analysis revealed that age (adjusted odds ratio [OR]: 1.10, 95% CI: 1.052-1.158), initial Glasgow Coma Scale (adjusted OR: 0.74, 95% CI: 0.647-0.837), and maximal CRP without surgical clipping (adjusted OR: 1.27, 95% CI: 1.066-1.516) were significantly associated with poor neurological outcomes in SAH patients. CONCLUSIONS: Early serial measurements of CRP may be used to predict neurological outcomes of SAH patients. Furthermore, maximal CRP levels within four days post-SAH are significantly correlated with poor neurological outcomes.

The differential neurologic prognosis of low-flow time according to the initial rhythm in patients who undergo extracorporeal cardiopulmonary resuscitation.

BACKGROUND: Limited data is available on the association between low-flow time and neurologic outcome according to the initial arrest rhythm in patients underwent extracorporeal cardiopulmonary resuscitation (ECPR). METHODS: Between September 2004 and December 2018, 294 patients with in-hospital cardiac arrest (IHCA) were included in this analysis. We classified the patients into asystole (n = 42), pulseless electrical activity (PEA, n = 163) and shockable rhythm (n = 89) according to their initial rhythm. Primary outcome was poor neurologic outcome defined as Cerebral Performance Categories scores of 3, 4, and 5. RESULTS: One-hundred ninety IHCA patients (64.6%) had poor neurologic outcomes. There was significantly worse neurologic outcomes among IHCA patients according to their initial rhythm (asystole [88.1%], PEA [66.3%], and shockable rhythm [50.6%], p < 0.001). The PEA group and the shockable rhythm group showed a significant association between low-flow time and neurologic outcomes while this relationship was not observed in the asystole group: PEA [ρ = 0.224, p = 0.005], shockable rhythm [ρ = 0.298, p = 0.006]), and asystole [ρ = -0.091, p = 0.590]. The best discriminative CPR to pump-on time for neurologic outcome was 22 min in the PEA group (area under the curve 0.687, 95% confidence interval [CI] 0.610-0.758, p < 0.001) and 46 min in the shockable rhythm group (area under the curve 0.671, 95% CI 0.593-0.743, p < 0.001). CONCLUSIONS: The effect of interplay between arrest rhythm and low-flow time might be helpful for decisions about team activation and management for ECPR and could provide information for early neurologic prognosis.

Safety and feasibility of ultrasound-guided placement of peripherally inserted central catheter performed by neurointensivist in neurosurgery intensive care unit.

We evaluated the safety and feasibility of ultrasound-guided peripherally-inserted central venous catheters (PICC) by a neurointensivist at the bedside compared to fluoroscopy-guided PICC and conventional central venous catheter (CCVC). This was a retrospective study of adult patients who underwent central line placement and were admitted to the neurosurgical intensive care unit (ICU) between January 2014 and March 2018. In this study, the primary endpoint was central line-induced complications. The secondary endpoint was initial success of central line placement. Placements of ultrasound-guided PICC and CCVC performed at the bedside if intra-hospital transport was inappropriate. Other patients underwent PICC placement at the interventional radiology suite under fluoroscopic guidance. A total of 191 patients underwent central line placement in the neurosurgery ICU during the study period. Requirement for central line infusion (56.0%) and difficult venous access (28.8%) were the most common reasons for central line placement. The basilic vein (39.3%) and the subclavian vein (35.1%) were the most common target veins among patients who underwent central line placement. The placements of ultrasound-guided PICC and CCVC at the bedside were more frequently performed in patients on mechanical ventilation (p = 0.001) and with hemodynamic instability (p <0.001) compared to the fluoroscopy-guided PICC placement. The initial success rate of central line placement was better in the fluoroscopy-guided PICC placement than in the placements of ultrasound-guided PICC and CCVC at the bedside (p = 0.004). However, all re-inserted central lines were successful. There was no significant difference in procedure time between the three groups. However, incidence of insertional injuries was higher in CCVC group compared to PICC groups (p = 0.038). Ultrasound-guided PICC placement by a neurointensivist may be safe and feasible compared to fluoroscopy-guided PICC placement by interventional radiologists and CCVC placement for neurocritically ill patients.

Outcomes of mechanical ventilation according to WIND classification in pediatric patients.

BACKGROUND: The outcomes of weaning processes are not well known in pediatric patients, and the International Conference Classification on weaning from mechanical ventilation showed limited application. We evaluate the relationship between the new Weaning according to a New Definition (WIND) classification and outcome in pediatric patients. METHODS: We conducted a retrospective cohort study in a tertiary pediatric intensive care unit (ICU). We included patients under 18 years of age who received invasive mechanical ventilation for more than 24 h and excluded cases with other than the first ICU admissions, tracheostomy with home ventilation before admission, intubation or weaning processes conducted in other ICU, and weaning with extracorporeal membrane oxygenation. Weaning processes were classified into four groups according to weaning duration after the first separation attempt (SA): no-SA, short weaning (< 24 h), difficult weaning (24 h-7 days), and prolonged weaning (> 7 days). Mortality rates were compared across groups using the Kruskal-Wallis test, and risk factors for the no-SA group were analyzed by multivariate logistic regression tests with age, sex, severity score at admission, admission type, and underlying disease as variables. RESULTS: Among 313 patients, 224 were enrolled and had a median age of 2.1 (interquartile range 0.5-6.6) years. Spontaneous breathing tests were done in 70.1% of enrolled patients. The median duration of intubation to the first SA was 4 (range 0-36) days, and 92.8% patients underwent the first SA within 14 days. The mortality rate was 0% in the short (0/99) and difficult (0/53) weaning groups and 17.9% (5/28) in the prolonged weaning group (p < 0.001). The mortality rate of the no-SA group was 93.2% (41/44). Admission severity (hazard ratio 1.036, confidence interval 1.022-1.050) and underlying oncologic disease (hazard ratio 7.341, confidence interval 3.008-17.916) were independent risk factors for lack of SA. CONCLUSIONS: In conclusion, WIND classification is associated with ICU mortality in pediatric patients. Further studies of this association are required to improve protocols associated with the weaning process and clinical outcomes. Trial registration Retrospectively registered.

Neurologic Outcomes in Patients Who Undergo Extracorporeal Cardiopulmonary Resuscitation.

BACKGROUND: This study aimed to develop a risk prediction model for neurologic outcomes in patients who underwent extracorporeal cardiopulmonary resuscitation (ECPR). METHODS: Between May 2004 and April 2016, a total of 274 patients who underwent ECPR were included in this analysis. The primary outcome was neurologic status on discharge from the hospital, as assessed by Cerebral Performance Categories (CPC) scale. To develop a new predictive scoring system, backward stepwise elimination and a z-score-based scoring scheme were used on the basis of logistic regression analyses. RESULTS: A total of 95 patients (34.7%) survived until discharge. Of these, 78 patients (28.5%) had favorable neurologic outcomes (CPC scores of 1 or 2). In the multivariable logistic regression analysis, significant predictors of poor neurologic outcome included age older than 65 years, initial Sequential Organ Failure Assessment score greater than 13 points, first monitored arrest rhythm, low-flow time longer than 30 minutes, initial pulse pressure less than 25 mm Hg, initial mean arterial pressure less than 70 mm Hg, and serum glucose level greater than 300 mg/dL. There was also a significant interaction between age and low-flow time. The newly developed neurologic outcome score after ECPR (nECPR) more effectively predicted poor neurologic outcome (C-statistic, 0.867; 95% confidence interval, 0.823 to 0.912) than the former ECPR score (p = 0.019) and the survival after venoarterial ECMO score (p < 0.001). CONCLUSIONS: The investigators created a risk prediction model for neurologic outcomes using independent predictors and the interaction between age and low-flow time, and this new scoring system could predict early neurologic prognosis more effectively in ECPR-treated patients. It may be help guide decisions in ECPR management for intensivists, cardiovascular surgeons, or cardiologists.

Early prediction of neurological outcome after barbiturate coma therapy in patients undergoing brain tumor surgery.

After a difficult brain tumor surgery, refractory intracranial hypertension (RICH) may occur due to residual tumor or post-operative complications such as hemorrhage, infarction, and aggravated brain edema. We investigated which predictors are associated with prognosis when using barbiturate coma therapy (BCT) as a second-tier therapy to control RICH after brain tumor surgery. The study included adult patients who underwent BCT after brain tumor surgery between January 2010 and December 2016. The primary outcome was neurological status upon hospital discharge, which was assessed using the Glasgow Outcome Scale (GOS). In the study period, 4,296 patients underwent brain tumor surgery in total. Of these patients, BCT was performed in 73 patients (1.7%). Among these 73 patients, 56 (76.7%) survived to discharge and 25 (34.2%) showed favorable neurological outcomes (GOS scores of 4 and 5). Invasive monitoring of intracranial pressure (ICP) was performed in 60 (82.2%) patients, and revealed that the maximal ICP within 6 h after BCT was significantly lower in patients with favorable neurological outcome as well as in survivors (p = 0.008 and p = 0.028, respectively). Uncontrolled RICH (ICP ≥ 22 mm Hg within 6 h of BCT) was an important predictor of mortality after BCT (adjusted hazard ratio 12.91, 95% confidence interval [CI] 2.788-59.749), and in particular, ICP ≥ 15 mm Hg within 6 h of BCT was associated with poor neurological outcome (adjusted odds ratio 9.36, 95% CI 1.664-52.614). Therefore, early-controlled ICP after BCT was associated with clinical prognosis. There were no significant differences in the complications associated with BCT between the two neurological outcome groups. No BCT-induced death was observed. The active and timely control of RICH may be beneficial for clinical outcomes in patients with RICH after brain tumor surgery.

The prognostic value of optic nerve sheath diameter in patients with subarachnoid hemorrhage.

BACKGROUND: We evaluated the role of optic nerve sheath diameter (ONSD) using brain computed tomography (CT) in predicting neurological outcomes of patients with subarachnoid hemorrhage (SAH). METHODS: This was a retrospective, multicenter, observational study of adult patients with SAH admitted between January 2012 and June 2017. Initial brain CT was performed within 12 h from onset of SAH, and follow-up brain CT was performed within 24 h from treatment of a ruptured aneurysm. Primary outcome was neurological status at 6-month follow-up assessed with the Glasgow Outcome Scale (GOS, 1 to 5). RESULTS: Among 223 SAH patients, 202 (90.6%) survived until discharge. Of these survivors, 186 (83.4%) manifested favorable neurological outcomes (GOS of 3, 4, or 5). In this study, the ONSDs in the group of patients with poor neurological outcome were significantly greater than those in the favorable neurological outcome group (all p < 0.01). Intracranial pressure (ICP) was monitored in 21 (9.4%) patients during the follow-up CT. A linear correlation existed between the average ONSD and ICP in simple correlation analysis (r = 0.525, p = 0.036). Analysis of the receiver  operating characteristic curve for prediction of poor neurological outcome showed that ONSD had considerable predictive value (C-statistics, 0.735 to 0.812). In addition, the performance of a composite of Hunt and Hess grade and ONSD was increasingly associated with poor neurological outcomes than the use of each marker alone. CONCLUSIONS: ONSD measured with CT may be used in combination with clinical grading scales to improve prognostic accuracy in SAH patients.

Target Temperature Management May Not Improve Clinical Outcomes of Extracorporeal Cardiopulmonary Resuscitation.

PURPOSE: Target temperature management (TTM) and extracorporeal cardiopulmonary resuscitation (ECPR) have been established as important interventions during cardiopulmonary arrest. However, the impact of combined TTM and ECPR on clinical outcomes has not been studied in detail. METHODS: We reviewed the records of 245 patients who received extracorporeal life support (ECLS) between January 2012 and June 2015. Exclusion criteria were as follows: Extracorporeal life support performed for reasons other than cardiac arrest, age less than 18 years, and death within 24 hours. A total of 101 patients were finally included in the study. Twenty-five patients underwent TTM, and 76 patients did not. RESULTS: The patients' mean age was 55 ± 16.7 years. The mean cardiac arrest time was 44.6 ± 33.5 minutes. There were 84 patients whose cardiac arrest was due to a cardiac cause (83.2%) and 79 patients with in-hospital cardiac arrest (78.2%). There was a significant difference in average body temperature during the first 24 hours following ECPR (33.4°C vs 35.6°C; P = .001). The overall favorable neurological outcome rate was 34% and hospital survival rate was 47%. There was no difference in favorable neurological outcomes and hospital survival between the TTM and non-TTM groups (P = .91 and .84, respectively). On multivariate analysis of neurological outcomes and hospital survival, TTM was not a significant prognostic factor. CONCLUSION: We did not observe any benefits of TTM in patients undergoing ECPR. Natural hypothermia or normothermia related to ECLS may explain this result. Further research is needed to understand the role of TTM in ECPR.