Lung-protective Ventilation in Patients with Brain Injury: A Multicenter Cross-sectional Study and Questionnaire Survey in China.

BACKGROUND: Over the years, the mechanical ventilation (MV) strategy has changed worldwide. The aim of the present study was to describe the ventilation practices, particularly lung-protective ventilation (LPV), among brain-injured patients in China.
METHODS: This study was a multicenter, 1-day, cross-sectional study in 47 Intensive Care Units (ICUs) across China. Mechanically ventilated patients (18 years and older) with brain injury in a participating ICU during the time of the study, including traumatic brain injury, stroke, postoperation with intracranial tumor, hypoxic-ischemic encephalopathy, intracranial infection, and idiopathic epilepsy, were enrolled. Demographic data, primary diagnoses, indications for MV, MV modes and settings, and prognoses on the 60th day were collected. Multivariable logistic analysis was used to assess factors that might affect the use of LPV.
RESULTS: A total of 104 patients were enrolled in the present study, 87 (83.7%) of whom were identified with severe brain injury based on a Glasgow Coma Scale ≤8 points. Synchronized intermittent mandatory ventilation (SIMV) was the most frequent ventilator mode, accounting for 46.2% of the entire cohort. The median tidal volume was set to 8.0 ml/kg (interquartile range [IQR], 7.0-8.9 ml/kg) of the predicted body weight; 50 (48.1%) patients received LPV. The median positive end-expiratory pressure (PEEP) was set to 5 cmH2O (IQR, 5-6 cmH2O). No PEEP values were higher than 10 cmH2O. Compared with partially mandatory ventilation, supportive and spontaneous ventilation practices were associated with LPV. There were no significant differences in mortality and MV duration between patients subjected to LPV and those were not.
CONCLUSIONS: Among brain-injured patients in China, SIMV was the most frequent ventilation mode. Nearly one-half of the brain-injured patients received LPV. Patients under supportive and spontaneous ventilation were more likely to receive LPV. TRIAL REGISTRATION: ClinicalTrials.org NCT02517073 https://clinicaltrials.gov/ct2/show/NCT02517073.

INTRODUCTION

Over the past two decades, lung-protective ventilation (LPV), including limiting the tidal volume (VT) and plateau pressure while providing adequate positive end-expiratory pressure (PEEP) levels, has gradually been adopted in the mechanical ventilation (MV) of patients with acute respiratory distress syndrome (ARDS).[123] Recently, several studies have also shown that LPV decreased the risk for the development of pulmonary complications in patients without ARDS.[4567] However, in patients with brain injury, particularly those with increased intracranial pressure, low VT and high PEEP might have potential deleterious effects on the cerebral perfusion.[89] Consequently, physicians are reluctant to apply the LPV in brain-injured patients.[1011] However, several studies have reported that high VT was associated with the development of ARDS in severe brain-injured patients[121314] or other patients without ARDS at the onset of MV.[1516] Sporadic reports suggested that LPV, as employed in brain-injured patients, might not significantly impair the intracranial hypertension[17] and might improve the clinical outcome and reduce the duration of MV.[1819] Roquilly et al. concluded that LPV was associated with a reduction in the duration of MV in patients with brain injury.[18] Protective ventilation could also improve clinical outcomes in patients after cardiac arrest.[19] To date, there is no consensus regarding the application of LPV among patients with brain injury. Until recently, only few studies have described the characteristics and practice of MV in patients with brain injury.[1011131419] We conducted a multicenter, 1-day point cross-sectional study to describe the ventilation practices, particularly LPV, among brain-injured patients in China.

METHODS

Study design

We conducted a multicenter, 1-day point, cross-sectional study on the practice of MV among patients with brain injury in China. The study was performed at 11:00 a.m. on August 10, 2015.

Ethics and dissemination

The study was approved by the Institutional Review Board of Beijing Tiantan Hospital, Affiliated with Capital Medical University. The Institutional Review Board specifically approved the informed consent waiver, reflecting the anonymous and purely observational nature of this study. During the study period, no attempt was made to change the routine clinical practice in each participating Intensive Care Unit (ICU).

Study population

Adult patients with brain injury, including traumatic brain injury, stroke (ischemic stroke, spontaneous intracerebral hemorrhage, and subarachnoid hemorrhage), postoperation with intracranial tumor, hypoxic-ischemic encephalopathy, intracranial infection and idiopathic epilepsy, and receiving MV for at least 24 h, were enrolled in the present study. Severe brain injury was identified using a Glasgow Coma Scale (GCS) ≤8 points.[20] The exclusion criteria included age <18 years, undergoing a spontaneous breathing trial with a T-piece, or participating in another MV trial during the study period.

Data collection

During the study period, the total number of patients in each ICU, the number of patients with brain injury, and the number of ventilated patients with brain injury were recorded. A uniform case report form was used to collect the data and was completed by the doctors in charge of the patient, who were provided with detailed instructions and the related definitions. For each patient enrolled, we collected baseline data, including demographics, the date of admission to the hospital and ICU, and the primary diagnosis. The severity of illness, as estimated using Sequential Organ Failure Assessment (SOFA) scores,[21] and the related clinical data were recorded upon admission to the ICU and at study entry. The GCS was recorded upon admission to the ICU and at study entry. The cases of organ failure (cardiovascular, respiratory, renal, hepatic, and hematologic), defined as more than 2 points according to SOFA scores,[21] were collected at admission to ICU. The date and reasons for the initiation of MV and mode of artificial airway were recorded. For each patient, the ventilator mode and settings and arterial blood gas analysis, including arterial pH, partial pressure of oxygen (PaO2)/fraction of inspired oxygen (FiO2) ratio (P/F ratio), and partial pressure of carbon dioxide partial pressure (PaCO2), were recorded under the settings described above. For patients receiving assist/control ventilation (A/C), volume-controlled ventilation (VCV), pressure-controlled ventilation (PCV), and pressure-regulated volume control ventilation (PRVC), the VT under mandatory ventilation was adapted; for patients receiving biphasic positive airway pressure (BIPAP), the VT under the high-level pressure was recorded; for patients receiving pressure support ventilation (PSV) and continuous positive airway pressure (CPAP), the VT of monitoring was adapted; and for patients receiving synchronized intermittent mandatory ventilation (SIMV) combined with PSV, the VT under mandatory ventilation and pressure support were both recorded and the larger of the two values was adapted. Based on the proportion of mandatory ventilation, we defined three MV mode groups: supported and spontaneous ventilator mode, including PSV and CPAP, was identified as Group 1; partially ventilator mode, including SIMV, A/C, BIPAP, and PRVC, was identified as Group 2; and totally mandatory ventilator mode, including PCV and VCV, was identified as Group 3.

LPV was defined as VT≤6 ml/kg of the predicted body weight (PBW) or VT≤8 ml/kg of the PBW and peak airway pressure <30 cmH2O.[1] Sepsis and septic shock were defined according to the American College of Chest Physicians/Society of Critical Care Medicine consensus conference definitions.[22] ARDS was identified according to the berlin definition.[23] The patients were followed up until hospital discharge, death, or 60 days after the day of investigation whichever occurred first. The date of discharge from ICU, prognosis of artificial airway, the Glasgow Outcome Scale (GOS) on the 60th day, duration of MV, and the weaning outcome were documented. On the afternoon of August 10, 2015, a questionnaire concerning the preferred MV mode and settings was sent to the physicians involved in the study, which was anonymous and collected by the investigator at each center. The details are presented in Supplement File 1.

Questionnaire of mechanical ventilation practice among patients with brain injury

Click here for additional data file.

Statistical analysis

Continuous variables were presented as median and interquartile range (IQR) values and compared using the Mann–Whitney U-test or Kruskal–Wallis test with Bonferroni's correction. Bonferroni's correction provides a straightforward approach to control the Type I error rate when multiple testing is performed. The adjusted alpha (α) level after Bonferroni's correction was shown as α’. The categorical variables were reported as numbers and percentages and subsequently compared using either Chi-square test or Fisher's exact test when appropriate. Multivariable logistic analysis was used to assess factors that might affect the use of LPV. All variables with a P < 0.2 and those which we thought might be associated with the use of LPV were included in the multivariate model. Backward selection based on the likelihood ratio test was used to select the final multivariate model for factors associated with the use of LPV. All analyses were performed using the statistical software package SPSS 21 (SPSS Inc., Chicago, IL, USA). A P < 0.05 was considered statistically significant.

RESULTS

We identified 47 ICUs in China, 20 of which were Neurologic ICUs (NICUs). The median number of beds in each ICU was 20 (IQR, 15–26). There were 839 patients in the ICUs of participating hospitals at the time of the present study and the bed occupancy rate was 79.8%; 365 patients were identified with brain injury, and the patient flowchart is shown in Figure 1. A total of 104 patients were enrolled in the study, including 32 patients with traumatic brain injury, 45 with stroke, 11 with hypoxic-ischemic encephalopathy, and 16 with other brain injuries. Moreover, 87 (83.7%) patients were identified with severe brain injury with GCS ≤8 points. The characteristics of the entire cohort were shown in Table 1 and the outcome of patients alive at ICU discharge was shown in Table 2.

Figure 1

Flowchart of patients into the study. ICU: Intensive Care Unit; MV: Mechanical ventilation; SBT: Spontaneous breathing trial; NPPV: Noninvasive positive pressure ventilation.

Table 1

Baseline characteristics of the patients in this study

Characteristics of all subjects	Traumatic brain injury (n = 32)	Stroke (n = 45)	Hypoxic-ischemic encephalopathy (n = 11)	Others* (n = 16)	Total (n = 104)
Age, median (IQR), years	51 (40, 68)	58 (48, 69)	47 (37, 84)	55 (37, 63)	55 (46, 67)
Male, n (%)	28 (87.5)	37 (82.2)	7 (63.6)	5 (31.3)	77 (74.0)
Height, median (IQR), cm	173.0 (170.0, 175.0)	172.0 (168.0, 175.0)	171.0 (157.0, 175.0)	161.5 (156.3, 171.0)	172.0 (165.0, 175.0)
SOFA score, median (IQR)	8 (6, 9)	7 (5, 9)	12 (10, 12)	7 (4, 10)	8 (6, 10)
GCS, median (IQR)	5 (3, 8)	5 (3, 6)	5 (3, 8)	7 (4, 9)	5 (3, 7)
GCS ≤8, n (%)	27 (84.4)	39 (86.7)	9 (81.8)	12 (75.0)	87 (83.7)
Organ failure, n (%)	17 (53.1)	28 (62.2)	9 (81.8)	9 (56.3)	63 (60.6)
Arterial blood gas analysis
pH, median (IQR)	7.45 (7.40, 7.48)	7.46 (7.43, 7.50)	7.41 (7.36, 7.47)	7.47 (7.43, 7.50)	7.46 (7.42, 7.49)
PaCO2, median (IQR), mmHg	35.5 (31.1, 41.8)	33.1 (29.2, 40.3)	36 (29.0, 40.0)	35.5 (30.3, 37.9)	35.0 (30.0, 40.4)
PaO2/FiO2 ratio, median (IQR)	312.5 (214.6, 392.4)	262.5 (178.5, 355)	280.0 (128.3, 548.6)	312.8 (231.9, 373.2)	286.3 (201.0, 372.2)
ICU mortality, n (%)	6 (18.8)	8 (17.8)	3 (27.3)	1 (6.3)	18 (17.3)
GOS on the 60th day, median (IQR)	3 (2, 3)	2 (2, 3)	2 (1, 4)	2 (2, 3)	3 (2, 3)
Tracheostomy, n (%)	19 (59.4)	27 (60.0)	7 (63.6)	10 (62.5)	63 (60.6)

*Other patients included 16 patients, 9 with brain tumor, 3 with central nervous system involved were identified as Sjogren syndrome, drug abuse, and leukoencephalopathy, respectively, two diagnosed as idiopathic epilepsyǀ, and two identified as intracranial infection. IQR: Interquartile range; SOFA: Sequential Organ Failure Assessment; GCS, Glasgow Coma Scale; ICU: Intensive Care Unit; GOS: Glasgow Outcome Scale; PaCO2: Partial pressure of carbon dioxide partial pressure; PaO2: Partial pressure of oxygen; FiO2: Fraction of inspired oxygen.

Table 2

Outcome of patients alive at ICU discharge

Outcome	Traumatic brain injury (n = 26)	Stroke (n = 37)	Hypoxic-ischemic encephalopathy (n = 8)	Others* (n = 15)	Total (n = 86)
MV duration, median (IQR), days	11 (8, 23)	12 (6, 21)	15 (10, 71)	21 (7, 30)	13 (7, 25)
LOS in ICU, median (IQR), days	23 (11, 30)	20 (15, 35)	18 (14, 69)	23 (11, 47)	21 (14, 34)
Patients with weaning, n (%)	24 (92.3)	29 (78.4)	4 (50.0)	8 (53.3)	65 (75.6)

*Other patients included 15 patients, 8 with brain tumor, 3 with central nervous system involved were identified as Sjogren syndrome, drug abuse, and leukoencephalopathy, respectively, two diagnosed as idiopathic epilepsy, and two identified as intracranial infection. IQR: Interquartile range; ICU: Intensive Care Unit; MV: Mechanical ventilation; LOS: Length of stay.

Indication for mechanical ventilation and artificial airway

In the present study, the most frequent indication for MV was acute respiratory failure, accounting for 76% of the entire cohort; pneumonia was the most common indication (29.8%), followed by postoperative respiratory failure (21.2%), aspiration (14.4%), trauma (10.6%), sepsis (5.8%), and heart failure (3.8%). Among the patients, 10 (9.6%) were identified with ARDS. There were overlaps among pneumonia, aspiration, trauma, heart failure, and sepsis. An abnormal respiratory rhythm was the second common indication for MV, accounting for 24% of the entire cohort. A total of 66 (63.5%) patients were ventilated using an endotracheal tube and 38 (36.5%) patients were ventilated using a tracheostomy tube. Among the patients who were ventilated with endotracheal tubes, 93.9% tubes were passed through the mouth and 6.1% of the tubes were passed through the nose. Among the 38 patients who were ventilated with tracheostomy tubes, the tracheostomy was performed at a median of 2 days (IQR, 0–9 days) after intubation.

Ventilation mode and settings

Among all 104 patients, 48 (46.2%) received SIMV combined with PSV (volume-controlled SIMV 28.8%, pressure-controlled SIMV 17.3%), which was the most frequent mode of ventilation, followed by PSV (18.3%), CPAP (10.6%), A/C (9.6%), PCV (4.8%), VCV (4.8%), BIPAP (2.9%), and PRVC (2.9%). Although mandatory ventilation was the most common method used, the distribution of the ventilator mode was different among patients with different brain injuries (P = 0.035).

The median VT of all patients was set to 8.0 ml/kg (IQR, 7.0–8.9) of the PBW. There was no significant difference among different brain injuries (P = 0.729). The details of the ventilator modes and settings among patients with different brain injuries are presented in Table 3. The VT of patients under partial mandatory ventilation was slightly higher than that of patients receiving supported and spontaneous ventilation (8.3 ml/kg [IQR, 7.5–9.0] vs. 7.4 ml/kg [IQR, 6.4–8.4], P = 0.025), accompanied by a higher peak pressure (20 cmH2O [IQR, 17–24] vs. 15 cmH2O [IQR, 13–18], P = 0.001). Compared with patients receiving supported and spontaneous ventilation, the peak pressure under total mandatory ventilation was higher (15 cmH2O [IQR, 13–18] vs. 21 cmH2O [IQR, 18.0–24.5], P = 0.005). The details of the ventilator data among different ventilator modes are shown in Table 4. Compared with patients without ARDS, the VT was slightly lower among patients with ARDS although this value was not statistically significant (8.0 ml/kg [IQR, 7.2–8.9] vs. 7.3 ml/kg [IQR, 6.2–8.0], P = 0.073).

Table 3

Ventilator modes and settings among patients with different brain injury

Variables	Traumatic brain injury (n = 32)	Stroke (n = 45)	Hypoxic-ischemic encephalopathy (n = 11)	Others* (n = 16)	P
Mode, n (%)
Group 1	5 (15.6)	17 (37.8)	1 (9.1)	7 (43.8)	0.035†
Group 2	25 (78.1)	25 (55.6)	8 (72.7)	6 (37.5)
Group 3	2 (6.3)	3 (6.7)	2 (18.2)	3 (18.8)
Ventilator data
VT, median (IQR), ml/kg	8.0 (6.9, 8.9)	8.0 (7.2, 8.7)	8.3 (6.3, 8.7)	8.0 (7.3, 9.3)	0.729
PEEP, median (IQR), cmH2O	5 (4, 5)	5 (5, 6)	6 (5, 7)	5 (4, 7)	0.009‡

*Other patients included nine patients with brain tumor, three with central nervous system involved were identified as Sjogren syndrome, drug abuse, and leukoencephalopathy, respectively, two were diagnosed as idiopathic epilepsy and two were identified as intracranial infection; Group 1 of mode included PSV and CPAP; Group 2 of mode included SIMV, A/C ventilation, PRVC and BIPAP; Group 3 of mode included PCV and VCV. VT: Tidal volume; IQR: Interquartile range; PEEP: Positive end-expiratory pressure. The adjusted alpha level (α’) was 0.008 after Bonferroni's correction. †There was no significant difference in ventilation mode between patients with any two kinds of brain injuries, after Bonferroni's correction (Pmin = 0.019>α’). ‡There were significant differences in PEEP level between patients with traumatic brain injury and those with hypoxic-ischemic encephalopathy, even after Bonferroni's correction (P = 0.001<α’). PCV: Pressure-controlled ventilation; VCV: Volume-controlled ventilation; SIMV: Synchronized intermittent mandatory ventilation; A/C: Assist/control; PRVC: Pressure-regulated volume control; BIPAP: Biphasic positive airway pressure.

Table 4

Comparisons of ventilator data among different ventilation modes

Variables	Group 1 (n = 30)	Group 2 (n = 64)	Group 3 (n = 10)	P
VT (ml/kg)	7.4 (6.4, 8.4)	8.3 (7.5, 9.0)	7.6 (6.7, 8.3)	0.038*
PEEP (cmH2O)	5 (4, 5)	5 (5, 6)	5 (5, 6)	0.195
RR of setting (breaths/min)	NA	15 (12, 16)	15 (10, 15)	0.352
RR of monitoring (breaths/min)	20 (15, 22)	16 (13, 20)	15 (12, 15)	0.035†
pH	7.47 (7.44, 7.50)	7.45 (7.41, 7.49)	7.45 (7.26, 7.47)	0.053
PaCO2 (mmHg)	34.0 (29.8, 39.3)	34.5 (29.0, 39.8)	40.8 (33.8, 47.8)	0.095
Ratio PaO2/FiO2	265.6 (199.9, 357)	295 (202.6, 392.4)	326.1 (205.4, 373.8)	0.522
Peak pressure (cmH2O)	15 (13, 18)	20 (17, 24)	21 (18, 25)	0.001‡
Plateau pressure (cmH2O)	NA	12 (9, 15)	14 (8, 17)	0.660

Data are shown as median (IQR). Group 1 of mode included PSV and CPAP; Group 2 of mode included SIMV, A/C ventilation, PRVC and BIPAP; Group 3 of mode included PCV and VCV. VT: Tidal volume; PEEP: Positive end-expiratory pressure; RR: Respiratory rate. The adjusted alpha level (α’) was 0.017 after Bonferroni's correction. *There was no significant difference in tidal volume between any two groups, after Bonferroni's correction (Pmin = 0.025>α’); †There was no significant difference in respiratory rate between any two groups, after Bonferroni's correction (Pmin = 0.021>α’); ‡There were significant differences in Peak pressure between Groups 1 and 2, as well as Groups 1 and 3, even after Bonferroni's correction (P = 0.001<α’, P = 0.005<α’, respectively). IQR: Interquartile range; CPAP: Continuous positive airway pressure; PCV: Pressure-controlled ventilation; VCV: Volume-controlled ventilation; SIMV: Synchronized intermittent mandatory ventilation; A/C: Assist/control; PRVC: Pressure-regulated volume control; BIPAP: Biphasic positive airway pressure; PaCO2: Partial pressure of carbon dioxide partial pressure; PaO2: Partial pressure of oxygen; FiO2: Fraction of inspired oxygen. NA: Not available.

The median PEEP level was 5 cmH2O (IQR, 5–6). All PEEP values were no more than 10 cmH2O. Patients with hypoxic-ischemic encephalopathy received a higher level of PEEP than did those with traumatic brain injury (6 cmH2O [IQR, 5–7] vs. 5 cmH2O [IQR, 4–5], P = 0.001). For patients with severe hypoxemia (PaO2/FiO2<100 mmHg), moderate hypoxemia (100≤ PaO2/FiO2<200 mmHg), mild hypoxemia (200≤ PaO2/FiO2<300 mmHg), and P/F ratios higher than 300 mmHg, the PEEP levels were set to 8 cmH2O (IQR, 7–9), 5 cmH2O (IQR, 5–7), 5 cmH2O (IQR, 4–6), and 5 cmH2O (IQR, 4–5), respectively and there was a significant difference among different levels of hypoxemia (P = 0.005). Patients with severe hypoxemia had higher PEEP levels than the others did. In contrast, there was no significant difference among the other groups. A higher level of PEEP was applied for patients with ARDS than for non-ARDS patients (6.6 cmH2O [IQR, 5–8] vs. 5.0 cmH2O [IQR, 4–5], P < 0.05).

Lung-protective ventilation practice

A total of 50 (48.1%) patients received LPV. The comparison between patients who received LPV and those who did not revealed that there was no significant difference in baseline characteristics, ventilator mode, PEEP setting, arterial pH, P/F ratio, or PaCO2. It seemed that patients who received LPV had a slightly higher respiratory rate (RR) than those who did not (18 breaths/min [IQR, 15–22] vs. 16 breaths/min [IQR, 12–19], P = 0.02]. The distribution of the ventilator mode between LPV and non-LPV was different (P = 0.01). The details are shown in Table 5. According to the multivariable logistic regression analysis (covariates entered: age, gender, height, SOFA and GCS at admission to ICU, arterial pH, P/F ratio, PaCO2, ventilator mode, PEEP setting, RR of monitoring, and ICU category), age and gender were associated with the practice of LPV; compared with partially mandatory ventilation (Group 2), supported and spontaneous ventilation (Group 1) was associated with the use of LPV (odds ratio 5.401, 95% confidence interval 1.878–15.536, P = 0.002). The details are shown in Table 6. Among older patients (≥60 years), 26 (59.1%) individuals received LPV, but among younger patients (<60 years), 24 (40%) individuals received LPV, suggesting that older patients were more likely to receive LPV although there was no significant difference (P = 0.054).

Table 5

Comparison between patients received lung-protective ventilation and those did not

Variables	LPV (n = 50)	Non-LPV (n = 54)	P
Age, median (IQR), years	60 (47, 69)	52 (44, 65)	0.164
Male, n (%)	40 (80.0)	37 (68.5)	0.182
Height, mean (range), cm	172.5 (165, 175)	170 (160, 175)	0.073
SOFA score, median (IQR)	7 (6, 9)	8 (6, 10)	0.441
GCS, median (IQR)	5 (4, 7)	5 (3, 7)	0.175
pH, median (IQR)	7.46 (7.43, 7.48)	7.45 (7.41, 7.49)	0.523
PaCO2, median (IQR), mmHg	35.4 (31.0, 40.0)	34.5 (29.0, 41.0)	0.410
Ratio PaO2/FiO2, median (IQR)	273.1 (197.3, 360.3)	296.3 (221.9, 405.0)	0.081
Patients in NICU, n (%)	27 (54.0)	31 (57.4)	0.727
Brain injury category			0.573
Traumatic brain injury, n (%)	14 (28.0)	18 (33.3)
Stroke, n (%)	25 (50)	20 (37)
Hypoxic-ischemic encephalopathy, n (%)	5 (10.0)	6 (11.1)
Others*, n (%)	6 (12.0)	10 (18.5)
Ventilator mode, n (%)			0.010
Group 1	20 (40.0)	10 (18.5)
Group 2	24 (48.0)	40 (74.1)
Group 3	6 (12.0)	4 (7.4)
RR of monitoring, median (IQR), breaths/min	18 (15, 22)	16 (12, 19)	0.020
PEEP, median (IQR), mmHg	5.0 (4.8, 6.0)	5.0 (4.8, 6.0)	0.585
LOS in ICU, median (IQR), days	21.0 (12.8, 34.0)	18.0 (10.8, 28.5)	0.309
Mortality, n (%)	9 (14.5)	9 (21.4)	0.432
MV duration, median (IQR), days	14.0 (7.5, 25.5)	12.0 (7.0, 24.0)	0.537

*Other patients included 16 patients, 9 with brain tumor, 3 with central nervous system involved were identified as Sjogren syndrome, drug abuse and leukoencephalopathy, respectively, 2 diagnosed as idiopathic epilepsy, and two identified as intracranial infection. LPV: Lung-protective ventilation; IQR: Interquartile range; SOFA: Sequential Organ Failure Assessment; GCS: Glasgow Coma Scale; NICU: Neurologic Intensive Care Unit; Group 1 of mode included PSV and CPAP; Group 2 of mode included SIMV, A/C ventilation, PRVC ventilation and BIPAP; Group 3 of mode included PCV and VCV; RR: Respiratory rate; PEEP: Positive end-expiratory pressure; LOS: Length of stay; MV: Mechanical ventilation. PCV: Pressure-controlled ventilation; VCV: Volume-controlled ventilation; SIMV: Synchronized intermittent mandatory ventilation; A/C: Assist/control; PRVC: Pressure-regulated volume control; BIPAP: Biphasic positive airway pressure; CPAP: Continuous positive airway pressure; PaCO2: Partial pressure of carbon dioxide partial pressure; PaO2: Partial pressure of oxygen; FiO2: Fraction of inspired oxygen.

Table 6

Association of factors with the use of lung-protective ventilation by multivariable logistic regression analysis

Variables	OR (95% CI)	P
Age	1.030 (1.001–1.059)	0.039
Gender	2.847 (1.014–7.995)	0.047
Ventilator mode
Group 1	5.401 (1.878–15.536)	0.002
Group 3	4.284 (0.934–19.654)	0.095

The model is adjusted for height, SOFA and GCS at admission to ICU, arterial pH, PaO2/FiO2, PaCO2, ventilator mode, PEEP, RR of monitoring and ICU category. CI: Confidence interval; Group 1 of mode included PSV and CPAP; Group 3 of mode included PCV and VCV. SOFA: Sequential Organ Failure Assessment; GCS: Glasgow Coma Scale; ICU: Intensive Care Unit; PEEP: Positive end-expiratory pressure; RR: Respiratory rate; PSV: Pressure support ventilation; CPAP: Continuous positive airway pressure; PCV: Pressure-controlled ventilation; VCV: Volume-controlled ventilation; OR: Odds ratio; PaCO2: Partial pressure of carbon dioxide partial pressure; PaO2: Partial pressure of oxygen; FiO2: Fraction of inspired oxygen.

Outcomes

There was no significant difference in ICU mortality among patients with different brain injuries (P = 0.548). The median GOS of the entire cohort was 3 points (IQR, 2–3), suggesting that most ventilated patients with brain injury could not live alone. A total of 63 (60.6%) patients received tracheostomy during their stay in the ICU. For patients who were alive upon discharge from the ICU, the median MV duration was 13 days (IQR, 7–25 days), the median length of stay in the ICU was 21 days (IQR, 14–34 days), and there was no difference among patients with different brain injuries in either MV duration or length of ICU stay (P = 0.530, P = 0.701, respectively). The details are shown in Table 2.

Questionnaire for physicians

On the afternoon of August 10, 2015, a questionnaire of preferred ventilator mode and settings for patients with brain injury was sent to the doctors in charge of the patients enrolled in the present study. Three questionnaires were delivered to each center; 25 questionnaires were not completed and 116 (82.3%) questionnaires were analyzed. Consistent with the results of the present study, SIMV (60.3%) was the most frequent ventilator mode, followed by PCV (13.8%), VCV (10.3%), PSV (8.6%), and others (7%). For neurologic patients, regardless of the presence of ARDS, the VT of 6–8 ml/kg was preferred to 8–10 ml/kg (58.6% vs. 41.4%, P = 0.012). There were more physicians in the general ICU preferring to practice LPV, compared with those in NICU (64.5% vs. 47.5%, P = 0.078), although the difference was not significant. A total of 51 (44%) doctors preferred a relatively lower PEEP level of <5 cmH2O whereas other physicians preferred to set the PEEP level between 5 and 10 cmH2O. None of the physicians set the PEEP level above 10 cmH2O. For patients with severe refractory hypoxemia, the recruitment maneuver was the most frequently used, followed by the prone position. Extracorporeal membrane oxygenation was rarely used.

DISCUSSION

In this multicenter, 1-day, cross-sectional study, MV mode, ventilator data, and the practice of LPV among brain-injured patients in the ICUs of China were described. To the best of our knowledge, studies describing the practice of LPV among neurologic patients were rare.[171819]

In the present study, the most frequent indication for MV was acute respiratory failure, similar to nonneurologic patients, but abnormal respiratory rhythm and aspiration were more common.[2425] Perhaps further attention should be devoted to the consciousness and airway protection of patients with brain injury. During the study period, 36.5% of the patients were ventilated using a tracheostomy tube, which was slightly higher than the number of patients treated in the international study by Esteban et al., and the tracheostomy was performed at a median of 2 days (IQR, 0–9) after MV, in contrast to the treatment for 11 days (IQR, 5–19) described by Esteban et al.[25] For patients with brain injury, physician preferred early tracheostomy (performed within 10 days after initiation of laryngeal intubation). However, whether early tracheostomy could reduce the duration of MV, the incidence of ventilation-associated pneumonia and hospital mortality in critically ill patients remains controversial.[2627]

In the present study, SIMV was the most frequent ventilator mode for patients with brain injury. In contrast to the recent studies, A/C was the most common and SIMV dramatically declined.[311] More patients received LPV than in the study of Kahn et al. among patients with subarachnoid hemorrhage and identified as having acute lung injury (48.1% vs. 30.0%).[28] According to the results of the multivariable logistic analysis, older and male patients were associated with the use of LPV. There was a significant difference in height between males and females (175 cm [IQR, 170–175] vs. 160 cm [IQR, 155–160], P < 0.001). That is, taller patients with higher PBW, as calculated based on the height, were more likely to receive LPV, which likely indicates that the LPV was not deliberately set but varied in accordance with the height of the patient. Compared with partially mandatory ventilation, supported and spontaneous ventilation was associated with the use of LPV. Patients receiving supportive and spontaneous ventilation were at the later stages of MV compared with those receiving mandatory ventilation during the present study (7.5 days [IQR, 3.0–18.5] vs. 4.5 days [IQR, 2.0–11.5], P = 0.066). Among the thirty patients receiving supportive and spontaneous ventilation, 19 were ventilated with PSV with a median supportive pressure of 5 cmH2O [IQR, 4–6] and 11 were ventilated with CPAP. We speculated that supportive and spontaneous MV might be performed during the late stages of MV and that the supportive level was lowered to initiate weaning. Therefore, the VT was lower in patients receiving supportive MV. The results of the present study indicated that the use of LPV in the enrolled patients might represent the characteristics of the patient without a deliberate setting by physicians.

Between patients receiving LPV and those who did not, there was no significant difference in the P/F ratio or PaCO2 or in the outcomes, including ICU mortality, length of stay in ICU, and MV duration, in the present study. Thus, we concluded that LPV might be safe for patients with brain injury. Although a relatively higher PEEP was applied for patients with severe hypoxemia, the PEEP level applied in neurologic patients was lower than that in nonneurologic patients in a previous study.[11] First, the median P/F ratio of patients in the present study was 250 mmHg higher, indicating a relatively improved lung condition, which did not require a high level of PEEP. Second, considering that the reduction of cerebral blood flow might be caused by high PEEP levels, physicians prefer relatively lower PEEP levels,[89] consistent with the results of the questionnaire. The median VT was similar to the results of the study in Poland[24] but was lower than that reported in the previous study.[25]

There are several limitations in this study. First, this study was a 1-day point study, which could not represent the entire cohort of ventilated patients with brain injury. Second, only 104 patients were enrolled in the present study and the small sample might affect the results. Third, we did not collect the data for nonbrain-injured patients in participating ICUs during the study period. Moreover, we did not record the comorbidity of the enrolled patients, which might affect the evolution of the disease severity; thus, the relationship between hospital mortality and ventilator settings, including VT and PEEP, was not analyzed. Furthermore, we did not record the lung complication, such as ventilator-associated pneumonia and pneumothorax, or the use of sedative and analgesic drugs, which might also affect the MV mode.

In conclusion, among brain-injured patients in China, SIMV was the most frequent ventilation mode. Nearly one-half of brain-injured patients received LPV. Patients under supportive and spontaneous ventilation were more likely to receive LPV. In the present study, the use of LPV in the enrolled patients might represent the characteristics of the patients not the deliberately settings of physicians.

Supplementary information is linked to the online version of the paper on the Chinese Medical Journal website.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.