This was a prospective observational study performed in an ICU of a teaching hospital from 2016 to 2018. The study protocol was approved by the hospital’s institutional review board. Due to the observational nature of this study, informed consent was waived. Patients who were admitted to the ICU for NIV as a first-line intervention were enrolled, and patients younger than 18 years old were excluded. Patients who were treated with NIV after high-flow nasal cannula or invasive mechanical ventilation were also excluded.
NIV (BiPAP Vision or V60; Philips Respironics, Carlsbad, CA) was initiated by the attending physicians in relation to previously described indications [16,17]. In patients with chronic obstructive pulmonary disease (COPD), the indications for NIV were respiratory rate more than 25 breaths/min, PaCO2 more than 45 mmHg, pH less than 7.35, PaO2/FiO2 less than 200 mmHg, and vigorous activity of the accessory respiratory muscles. In patients with hypoxemic respiratory failure, the indications for NIV were clinical presentation of respiratory distress at rest (such as active contraction of the accessory inspiratory muscles or paradoxical abdominal motion) and PaO2/FiO2 less than 300 mmHg.
A face mask (ZS-MZ-A Face Mask, Shanghai Zhongshan Medical Technology Co., Shanghai, China) was the first choice to connect the patient to the ventilator. The size of the mask was chosen to fit the patients’ face type. The S/T mode was used in COPD patients or other patients with labored breathing. The initial inspiratory positive airway pressure was 8–10 cmH2O, which was gradually increased to reach a tidal volume of 6–8 mL/kg or the maximal tolerated level. The initial expiratory positive airway pressure was 4 cmH2O. This was gradually increased to counterbalance the intrinsic positive end expiratory pressure in COPD patients. In patients with hypoxemic respiratory failure, it was increased to maintain alveoli patency and to elevate the end expiratory lung volume. In patients with heart failure, continuous positive airway pressure was used. The fractional concentration of oxygen was set to reach a peripheral oxygen saturation of >92%.
At the beginning of the treatment, continuous use of NIV was encouraged. Once the patient recovered from respiratory failure, liberation from NIV was considered, following our hospital protocols . However, intubation was indicated if respiratory failure progressively deteriorated. The criteria for intubation were as follows: respiratory or cardiac arrest, failure to maintain PaO2/FiO2 more than 100, development of a condition necessitating intubation to protect the airway (coma or seizure disorder) or to manage copious tracheal secretions, inability to correct dyspnea, lack of improvement of signs of respiratory muscle fatigue, and hemodynamic instability without response to fluids and vasoactive agents [16,17]. NIV failure was defined as requirement for intubation.
Assessment of delirium
Delirium was screened using the Confusion Assessment Method for the ICU (CAM-ICU) every morning from NIV initiation to termination . The assessment was performed by trained researchers. Delirium was assessed based on the following four features: (1) fluctuation in mental status, (2) inattention, (3) disorganized thinking, and (4) altered consciousness. Delirium was diagnosed in the presence of features 1 and 2 and either feature 3 or 4. The level of consciousness was assessed using the Richmond Agitation-Sedation Scale (RASS) . RASS is a 10-point scale ranging from unarousable (–5 points) through calm (0 points) to combative (4 points). The three subtypes of delirium were defined as follows . Hyperactive delirium was defined as present in patients with all positive daily RASS scores (range, 1–4 points) associated with every positive CAM-ICU assessment. Hypoactive delirium was defined as present in patients with all neutral or negative daily RASS scores (range, 0 to –3 points) associated with every positive CAM-ICU assessment. Mixed delirium was defined as present in patients with daily RASS scores that included both positive values (range, 1–4 points) and neutral or negative values (range, 0 to –3 points) associated with every positive CAM-ICU assessment.
The primary outcome was the rate of NIV failure. The second outcome included ICU mortality, hospital mortality, duration of NIV, length of stay in the ICU, length of stay in the hospital, and delirium days. The data was collected before hospital discharge. If patient died during hospitalization, the data was collected at the day of death.
Statistical software (SPSS 17.0; IBM Corp., Armonk, NY) was used to analyze the data. Continuous variables are presented as means with standard deviations or medians and interquartile ranges (25–75%) when appropriate. The normality of data distribution was analyzed using the Shapiro–Wilk test. Differences between groups were analyzed using the Student’s t test or Mann–Whitney U test as appropriate. Categorical variables are reported as frequencies and percentages. Differences between groups were analyzed using the chi-square test or Fisher’s exact test. Variables with p < 0.1, as computed in univariate analyses, and other clinical meaningful variables were entered into forward stepwise multivariate logistic regression analyses to identify the association between delirium and poor outcomes (e.g., NIV failure, ICU mortality, and hospital mortality). In addition, Cox regression analyses were also performed to retest the association between delirium and poor outcomes. Because delirium can occur at any time during NIV intervention, it was considered to be a time-varying covariate. The probability that patients would remain on NIV and in the ICU was analyzed using Kaplan–Meier curves (log-rank test). The threshold for statistical significance was set to p < 0.05.