Trends in Prevalence and Prognosis in Subjects With Acute Chronic Respiratory Failure Treated With Noninvasive and/or Invasive Ventilation

Inclusion and Exclusion Criteria

Subjects were identified through our institutional computerized database. The study was approved by the hospital ethics committee (13-73). We included all patients aged > 18 y admitted between January 1, 1998 and December 31, 2012, for ARF treated with NIV and/or invasive mechanical ventilation (IMV) initiated in the ICU or in the emergency department, and suffering from one of the following chronic respiratory diseases: COPD, bilateral bronchiectasis, obesity, restrictive disease, diffuse interstitial lung disease (ILD), and asthma. When NIV or IMV was initiated in the emergency department, subjects were managed by ICU physicians and transferred quickly to the ICU. Patients with cystic fibrosis, patients with restrictive syndrome secondary to neuromuscular disease, postoperative patients, and patients who were ventilated for reason other than respiratory failure, such as stroke or septic shock not due to pneumonia, were not included in the study.

Definitions

COPD was diagnosed according to the clinical and radiological criteria defined by the American Thoracic Society.¹⁹ For bilateral bronchiectasis, medical records were reviewed to ensure that cough and sputum productions were reported by subjects for at least 5 y before the admission to ICU and that bilateral dilated and thickened airways were visible on the chest x-rays performed just before ICU admission and/or on those performed at the admission on ICU and confirmed by a high-resolution computed tomography scan performed before ICU admission. Obesity was considered in subjects with body mass index > 30 kg/m² and was defined as morbid when body mass index > 40 kg/m². Chronic restrictive pulmonary disease included deformations of the rib cage due to kyphoscoliosis and post-tuberculosis sequelae including extended calcifications of the pleura as previously defined.¹⁰ Regarding chronic diffuse ILD, the disease was considered chronic when diagnosed at least 6 months before admission to the ICU, and diffuse parenchymal lung disease was confirmed on high-resolution computed tomography scan. Based on the consensus classification of the idiopathic interstitial pneumonias,²⁰ the following clinical conditions were distinguished: idiopathic pulmonary fibrosis (IPF) and non-IPF ILD including collagen vascular diseases, stage IV sarcoidosis, pneumoconiosis, chronic hypersensitivity pneumonitis, and drug-induced ILD. Severe acute asthma was defined according to the usual criteria.²¹ Nocturnal hypoventilation was previously documented by sleep polysomnography.

Morbid obesity was always considered as a primary etiology of chronic respiratory disease. In obese subjects without morbid obesity, respiratory failure was considered as secondary to obesity if no other identified cause for respiratory failure was found. Consequently, for non-morbidly obese subjects suffering from COPD, COPD was the primary diagnosis. Type of respiratory support was considered for the entire ICU stay, and subjects were classified as treated with NIV only, treated with IMV only, or transitioned from NIV to IMV.

Noninvasive Ventilation

Our ICU is a mixed 21-bed ICU in a university hospital that admits ∼1,000 patients/year, mostly for medical conditions. NIV has been used in our ICU since 1990. No other facilities existed for the acute management of either groups of these patients outside the ICU until 2010. From 2010 onward, NIV could also be initiated in the emergency department. The usual criteria used in our ICU for admission were those used for the diagnosis of respiratory failure. The nurse-to-patient ratio increased from 1 to 3.5 to 1 to 2.5 in November 2010. Subjects were treated with ventilators that were not able to compensate for air leaks (Servo C and Servo 300 respirators, Siemens, Sölna, Sweden) until June 01, 2000. Thereafter, ventilators able to compensate for air leaks were used (Servo-i, Siemens; or Evita XL, Dräger, Telford, Pennsylvania). The decision to initiate NIV was made by the subject's attending physician according to the usual criteria used in our ICU, that is, neurologic signs of hypercapnia, signs of acute respiratory distress (dyspnea increased over usual, accessory muscle use, tachypnea, paradoxical breathing), and gas exchange abnormalities (S_aO2 < 90%, P_aCO2 > 45 mm Hg, pH < 7.35).

At NIV initiation, our main concern was to maximize the subject's compliance with therapy. For that, pressure support ventilation was started with low inspiratory pressure and gradually titrated upward as tolerated by the subjects. F_IO2 was set to obtain a S_pO2 of 90% or more, and PEEP was applied in case and control groups (typically 3–6 cm H₂O). The intubation decision was made by subject's attending physician, on the basis of respiratory or cardiac arrest, agitation or altered consciousness, ineffective cough, pneumothorax, or hemodynamic instability. In addition to these criteria, NIV failure was considered when ventilation was ineffective because of major air leaks, or in subjects whose respiratory state worsened under NIV. From 2011, NIV was considered for difficult-to-wean subjects to reduce the duration of IMV. Immediate NIV following extubation was applied in subjects who had increased P_aCO2 during the spontaneous breathing trial but who did not develop clinical signs of poor tolerance.

Data Collection

The following variables recorded on ICU admission were collected: age, sex, body mass index, prior treatment with corticosteroids, prior home oxygen therapy, prior diagnosis of sleep apnea syndrome, associated cardiac comorbidity for which subjects received treatment excepting hypertension, lung and/or upper airway cancer recently diagnosed or currently treated, P_aCO2 values before initiation of MV (recorded as ≥ 50 mm Hg), and the Simplified Acute Physiology Score (SAPS) II calculated within 24 h after admission. When diagnosed at admission, the following suspected causative factors for exacerbation of the disease leading to respiratory failure were noted: bronchitis or pneumonia, pneumothorax, and pulmonary embolism; these were then regrouped as precipitating factors. In addition, the following outcome variables were recorded: treatment with NIV, treatment with IMV, or treatment with corticosteroids. Only the first episode of ventilatory support for each subject was included. Regarding ICU outcomes, in addition to mortality, we recorded the duration of MV, the length of the ICU stay, and the proportion of subjects with ICU-acquired pneumonia, as previously defined.²²

Data Analysis

Qualitative or categorical variables were expressed as number and percentages and continuous variables as mean ± SD. We first examined temporal trends for changes in the frequency of COPD, bilateral bronchiectasis, ILD, obesity, restrictive disease, and asthma; second, we examined temporal trends for changes in NIV and IMV use and in mortality from 1998 to 2012 using the logistic regression analysis with time as covariate. Temporal trends in SAPS II scores, durations of MV, and ICU lengths of stay were assessed by linear regression. Then, subjects were compared as to whether they received NIV only, IMV only or whether they failed NIV. Finally, we examined 2 outcomes: acquisition of ventilator-associated pneumonia and in-ICU mortality. The data were compared using the chi-square test or Fisher's exact test for categorical variables and using a one-way analysis of variance test for continuous variables. Regarding multivariate analysis for mortality, variables with P ≤ 0.10 were entered into a logistic regression model. The results were expressed in terms of the odds ratio (OR) with corresponding 95% CI. Tests were 2-tailed, and a value of P < .05 was considered significant.