Noninvasive vs Conventional Mechanical Ventilation in Acute Respiratory Failure A Multicenter, Randomized Controlled Trial Teresa Honrubia, MD; Fernando J. Garcı´a Lo´pez, MD, MHS; Nieves Franco, MD; Margarita Mas, MD; Marcela Guevara, MD; Martı´n Daguerre, MD; Inmaculada Alı´a, MD; Alejandro Algora, MD; and Pedro Galdos, MD*; on Behalf of the EMVIRA Investigators†
Study objective: Noninvasive mechanical ventilation (NIMV) is beneficial for patients with acute respiratory failure (ARF) when added to medical treatment. However, its role as an alternative to conventional mechanical ventilation (CMV) remains controversial. Our aim was to compare the efficacy and resource consumption of NIMV against CMV in patients with ARF. Design: A randomized, multicenter, controlled trial. Setting: Seven multipurpose ICUs. Patients: Sixty-four patients with ARF from various causes who fulfilled criteria for mechanical ventilation. Intervention: The noninvasive group received ventilation through a face mask in pressuresupport mode plus positive end-expiratory pressure; the conventional group received ventilation through a tracheal tube. Measurements and results: Avoidance of intubation, mortality, and consumption of resources were the outcome variables. Thirty-one patients were assigned to the noninvasive group, and 33 were assigned to the conventional group. In the noninvasive group, 58% patients were intubated, vs 100% in the conventional group (relative risk reduction, 43%; p < 0.001). Stratification by type of ARF gave similar results. In the ICU, death occurred in 23% and 39% (p ⴝ 0.09) and complications occurred in 52% and 70% (p ⴝ 0.07) in the noninvasive and conventional groups, respectively. There were no differences in length of stay. The Therapeutic Intervention Score System-28, but not the direct nursing activity time, was lower in the noninvasive group during the first 3 days. Conclusions: NIMV reduces the need for intubation and therapeutic intervention in patients with ARF from different causes. There is a nonsignificant trend of reduction in ICUs and hospital mortality together with fewer complications during ICU stay. (CHEST 2005; 128:3916 –3924) Key words: intratracheal; intubation; mask; randomized controlled trial; respiration, artificial; respiratory insufficiency Abbreviations: APACHE ⫽ acute physiology and chronic health evaluation; ARF ⫽ acute respiratory failure; CI ⫽ confidence interval; CMV ⫽ conventional mechanical ventilation; EMVIRA ⫽ Ensayo de Modalidades de Ventilacio´n en Insuficiencia Respiratoria Aguda (Trial on Modalities of Ventilation in Acute Respiratory Failure); Fio2 ⫽ fraction of inspired oxygen; NIMV ⫽ noninvasive mechanical ventilation; PEEP ⫽ positive end-expiratory pressure; PS ⫽ pressure support; TISS-28 ⫽ Simplified Therapeutic Intervention Scoring System-28
30% of patients admitted to an A pproximately ICU receive mechanical ventilation to treat acute
respiratory failure (ARF), and ⬎ 90% of them receive ventilation through endotracheal intubation.1 However, the insertion of an artificial airway can lead to numerous side effects related to sedation, airway damage, and respiratory infections.2– 4 Numerous randomized trials and reviews5–18 have shown benefits from the use of noninvasive mechanical ventilation (NIMV) when compared with conventional medical treatment in the treatment of 3916
ARF. The benefit of NIMV in hypoxemic or hypercapnic ARF with criteria for endotracheal intubation as an alternative to conventional mechanical ventilation (CMV) has been widely studied in nonrandomized trials.19 –22 Only two randomized trials23,24 have assessed NIMV as an alternative to CMV in patients with ARF and criteria for endotracheal intubation. They showed NIMV to be as efficacious as CMV in improving gas exchange, reducing complications,23 and avoiding intubation without changes in mortality.24 As NIMV applied through a face mask or nasal Clinical Investigations
mask may avoid the need for tracheal intubation, it might offer the additional benefits of a lesser need for sedation and a decrease in the risk of airway damage or pneumonia.25 Accordingly, if one assumes that as complications reduce, so do morbidity and mortality, NIMV might improve the outcome of ARF patients who require mechanical ventilation. For editorial comment see page 3790 We aimed to compare the efficacy of NIMV with CMV in patients with ARF from different causes with criteria for mechanical ventilation. Efficacy was assessed with the avoidance of intubation, medical complications, and the consumption of resources.
Materials and Methods
intercostal muscles; 3 ⫽ vigorous activity of accessory muscles with contraction; and 4 ⫽ vigorous activity with contraction of accessory muscles and paradoxical abdominal breathing pattern. Once the patient fulfilled these criteria, the final decision to enroll was made by the attending clinician. A consistency analysis was performed between the above criteria for intubation and the decision to intubate made by a physician. The index was 0.76 when one of the criteria was either the Kelly scale of neurologic dysfunction or the scale of use of accessory respiratory muscles. The reasons for exclusion were the presence of any condition immediately threatening life (imminent respiratory arrest, deep coma [Kelly score of 6], shock), contraindications to NIMV (coma of an origin other than hypercapnia, facial or digestive tract surgery), previous treatment with any mode of ventilation during the same admission, and decision not to apply mechanical ventilation for other reasons. Intervention The randomization scheme was stratified by center and made by a computer-based, pseudorandom number generator. Allocations were issued using opaque, sealed, and numbered envelopes.
Design, Setting, Subjects
NIMV
A multicenter randomized trial was carried out in seven ICUs. These units had sufficient expertise in the use of NIMV measured in terms of both time of use and the number of patients receiving ventilation. Patients with ARF were randomized to receive either NIMV or endotracheal intubation and CMV. The study was approved by the ethics committees of the participant centers. Written informed consent was given by participants or proxy relatives. The criteria for eligibility were ARF, both hypoxemic and hypercapnic, from causes previously defined26 –28 that had deteriorated despite standard medical management and at least three of the following criteria: Pao2/fraction of inspired oxygen (Fio2) ratio ⱕ 170; respiratory rate ⱖ 35 breaths/min; blood pH ⬍ 7.30; a score between 3 and 5 on the Kelly scale of neurologic dysfunction29; and a score of three or more points on a modified scale of accessory respiratory muscle use22: 1 ⫽ no visible respiratory activity in the neck muscles; 2 ⫽ respiratory activity in the neck muscles without active contraction of supraclavicular or
NIMV was applied with face mask (ResMed Ltd/North Ryde, New South Wales, Australia; and Respironics; Murrysville, PA). Standard ICU respirators were used (Dra¨ger Evita; Dra¨ger Medical; Lu¨beck, Germany; and Siemens; Siemens AG; Munich, Germany). The ventilatory mode applied was pressure support (PS) plus positive end-expiratory pressure (PEEP) with increasing PS level adjustment to achieve a baseline tidal volume from 5 to 7 mL/kg. In the subset of hypoxemic patients, PEEP was administered between 5 cm H2O and 10 cm H2O to achieve oxygen saturation ⬎ 90% with an Fio2 ⬍ 0.6. In hypercapnic patients, PEEP was applied between 5 cm H2O and 7 cm H2O to compensate for auto-PEEP.30 The application was continuous for the first 6 h,6,7 and no limit was set on the duration provided no criteria for failure appeared. For weaning spontaneous respiration with oxygen for 30 to 60 min was tried and if successful intermittent disconnections were followed. NIMV was considered to be successful if the patient remained in spontaneous respiration for at least 48 h after the withdrawal of NIMV. The criteria for NIMV failure were as follows: (1) cardiorespiratory arrest or imminent vital risk; (2) hemodynamic impairment that required vasoactive drugs; (3) severe damage at the nose bridge; and (4) impairment of any of the physiologic or gasometric variables with any decrease in the Pao2/Fio2 ratio, increase in Paco2 ⬎ 5 mm Hg, a decrease in pH of at least 0.05, or any increase on the Kelly scale of neurologic dysfunction or the scale of use of accessory respiratory muscles.
*From Intensive Care (Drs. Honrubia, Franco, Mas, and Galdos), Hospital de Mo´stoles, Mo´stoles; Clinical Epidemiology (Drs. Garcia Lo´pez and Guevara), Hospital Puerta de Hierro, Madrid; Intensive Care (Dr. Daguerre), Hospital Prı´ncipe de Asturias, Alcala´ de Henares; Intensive Care (Dr. Alia), Hospital de Getafe, Getafe; and Intensive Care (Dr. Algora), Fundacio´n Hospital de Alcorco´n, Alcorco´n, Spain. †A list of EMVIRA Investigators is given in the Appendix. This work was performed at the following institutions: Hospital de Mo´stoles, Mo´stoles; Hospital Prı´ncipe de Asturias, Alcala´ de Henares; Hospital Universitario de Getafe, Getafe; Fundacio´n Hospital de Alcorco´n, Alcorco´n; Hospital Morales Meseguer, Murcia; Hospital General Yagu¨e, Burgos; and Hospital Severo Ochoa, Legane´s. Funded by the Fondo de Investigacio´n Sanitaria from the Ministry of Health and Consumption of Spain (FIS 99/0043). Manuscript received February 19, 2005; revision accepted July 11, 2005. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml). Correspondence to: Fernando J. Garcı´a Lo´pez, MD, MHS, Unidad de Epidemiologı´a Clı´nica, Hospital Universitario Puerta de Hierro, San Martı´n de Porres, 4, 28035 Madrid, Spain; e-mail:
[email protected] www.chestjournal.org
CMV CMV was administered following the usual practice in each ICU. Assist-control ventilation was applied in 28 of 33 patients. Sedation was administered in 32 of 33 patients. All patients were intubated orally. Other therapeutic interventions were determined by the attendant physician. The use of weaning protocols was required in advance. T-tube weaning and PS were equally used. Outcome Variables The primary outcome variable was intubation avoidance in the NIMV group compared with the CMV group. Secondary clinical CHEST / 128 / 6 / DECEMBER, 2005
3917
variables were mortality in the ICU and at hospital discharge, clinical complications previously defined in the study protocol,4,31,32 and total ventilation time. The lengths of ICU and hospital stay, and the Simplified Therapeutic Intervention Scoring System-28 (TISS-28)33 scale for 7 days were also measured. Seven measurements were made for the time of nursing activity, the first for 8 h immediately after random assignment, and the other six measurements for 4 h each in every subsequent work shift. Moreover, we assessed the predictive factors for the failure of NIMV. Potential predictive variables were the primary cause of ARF, APACHE (acute physiology and chronic health evaluation) II score, baseline blood gases, in particular Paco2, changes in breathing rate, Kelly score, scale of accessory respiratory muscle use, Pao2/Fio2 ratio, Paco2, and pH at the first hour. Statistical Analysis A sample of 60 patients per group, for an ␣ value of 0.05 and a power of 0.90, would allow the detection of a reduction in the percentage of intubation of up to 18% in the group assigned to NIMV compared with the CMV group, and a reduction in its length of stay of 47%. However, after 64 patients had been enrolled, the Steering Committee decided to stop the study because of a decrease in the rate of enrollment, which the Steering Committee ascribed to a preference of attending physicians to treat directly with NIMV. All analyses were made by assigned treatment. Proportions were compared using the Mantel-Haenszel test stratified by center. Mortality was analyzed using Kaplan-Meier curves and log-rank tests. Means were compared using analysis of variance or Mann-Whitney U test. Length-of-stay variables were transformed into their corresponding natural logarithm. The probability of
intubation in the NIMV group was analyzed with multiple logistic regression. In the analyses of intubation avoidance, hypothesis testing was one sided and the significance level was set at p ⬍ 0.025. In the other analyses, hypothesis testing was two sided, and the significance level was established at p ⬍ 0.05. Two-sided 95% confidence intervals (CIs) was used for all analyses.34
Results Between November 1999 and September 2001, 64 patients were recruited; 31 were assigned to NIMV, and 33 were assigned to CMV. All randomized patients were included in the statistical analysis; no patients were unavailable for follow-up. Information on the number of potentially eligible patients who were admitted was only obtained in five of the seven centers during the study period (Fig 1). The characteristics of the patients at enrollment are listed in Table 1. Concerning inclusion criteria, all patients but one fulfilled at least one of the scales (3 to 5 points in the Kelly scale, and at least 3 in the scale of use of accessory respiratory muscles). No normocapnic patient had a Kelly score ⬎ 3, and three patients had pH between 7.24 and 7.30. Although the number of patients with a previous diagnosis of COPD was quite similar in the two groups, there were a greater number of patients with acute exacerbations
Figure 1. Participant flow through each stage. 3918
Clinical Investigations
Table 1—Baseline Characteristics in Patients at Randomization* Characteristics
NIMV (n ⫽ 31)
CMV (n ⫽ 33)
Male gender Age, yr ICU stay prior to randomization, h Hospital stay prior to randomization, d Pao2/Fio2 ratio Pao2/Fio2 ratio ⱕ 170 Respiratory frequency, breaths/min Respiratory frequency ⱖ 35 breaths/min Paco2, mm Hg Paco2 ⬎ 45 mm Hg pH pH ⬍ 7.30 Three to 5 points on Kelly scale of neurologic dysfunction Scale of use of accessory respiratory muscles ⱖ 3 points APACHE II† Heart rate, beats/min Heart rate ⱖ 140 beats/min Systolic BP, mm Hg Diastolic BP, mm Hg Cause of acute respiratory failure Exacerbation of COPD Pneumonia Cardiogenic acute pulmonary edema Bronchospasm Other‡ Previous diagnosis of COPD
22 (71) 70 ⫾ 11 0.7 (0.1–2.7) 1 (0–2) 119 ⫾ 66 27 (84) 37 ⫾ 8 24 (77) 72 ⫾ 28 25 (81) 7.27 ⫾ 0.09 20 (65) 18 (58) 25 (81) 25 ⫾ 7 110 ⫾ 21 11 (35) 145 ⫾ 30 79 ⫾ 18
23 (70) 65 ⫾ 16 0.5 (0.2–6.0) 1 (0–3) 120 ⫾ 59 26 (82) 34 ⫾ 11 19 (58) 70 ⫾ 30 23 (70) 7.27 ⫾ 0.13 21 (64) 17 (51) 26 (79) 24 ⫾ 10 113 ⫾ 27 10 (30) 145 ⫾ 33 73 ⫾ 17
11 (35) 8 (26) 9 (29) 1 (3) 2 (7) 20 (65)
7 (21) 10 (30) 7 (21) 3 (9) 6 (18) 18 (55)
*Data are presented as No. (%), mean ⫾ SD, or median (25 to 75 percentile). †First 24 h after admission to the ICU. ‡In the NIMV group, tetraparesis associated with rhabdomyolisis and diffuse alveolar hemorrhage; in the CMV group, pulmonary embolism, fat embolism, ARDS secondary to acute pancreatitis, sepsis in COPD, obstructive sleep apnea syndrome, and chest trauma.
of COPD and acute pulmonary edema as the cause for ARF in the NIMV group. Endotracheal Intubation In the group assigned to NIMV, 18 patients (58%) were intubated, vs 33 patients (100%) in the group
assigned to CMV (relative risk reduction, 43% [95% CI, 22 to 58%; p ⫽ 0.00003]; number needed to treat, 2.4 [95% CI, 1.7 to 4.1]; Fig 2). The analysis further stratified by the presence or not of exacerbation of COPD or cardiogenic acute pulmonary edema gave similar results (relative risk reduction,
Figure 2. Effect of NIMV vs CMV on endotracheal intubation. www.chestjournal.org
CHEST / 128 / 6 / DECEMBER, 2005
3919
42%; 95% CI, 17 to 52%), as did stratification by ARF without previous COPD, ARF due to exacerbation of COPD, ARF with previous COPD due to other causes (relative risk reduction, 42%; 95% CI, 19 to 55%), and stratification by hypercapnia at enrollment (relative risk reduction, 38%; 95% CI, 15 to 54%). In a subgroup analysis, there were no differences in relative risk reduction in patients with or without a history of COPD. Of eight patients with pneumonia (four with and four without a history of COPD), NIMV failed in all patients (Fig 2). The reasons for intubation in NIMV patients are shown in Table 2. Mortality In the NIMV group, 7 patients died (23%) in the ICU, vs 13 patients (39%) from the CMV group (relative risk reduction, 47%; 95% CI, ⫺ 12 to 75%; p ⫽ 0.09). This tendency toward mortality reduction was homogeneous in all subgroups analyzed (Fig 3). There were no statistically significant differences in the ICU survival time between the two groups (p ⫽ 0.60) [Fig 4]. Of the 18 patients assigned to NIMV who required endotracheal intubation, 7 patients died (39%). All patients from the NIMV group who died in the ICU had been intubated. In the NIMV group, 10 patients (32%) died in the hospital, vs 14 patients (42%) in the CMV group (relative risk reduction, 24%; 95% CI, ⫺ 34 to 61%; p ⫽ 0.30). There were no statistically significant differences in hospital survival time between the two groups (p ⫽ 0.43). One patient died of acute exacerbation of chronic renal failure 21 days after ICU discharge without having been intubated. Of 18 patients with baseline diagnosis of pneumonia, 12 patients died in ICU, 9 from septic shock with acute renal failure and 3 from refractory ARF. The causes of death are listed in Table 3.
Table 2—Indications for Intubation in the NIMV Group Indications
No. (%)
Increase in Kelly score Impairment of ventilation* Cardiorespiratory arrest or imminent vital risk Impairment of oxygenation† Agitation Intolerance to face mask Transfer to other hospital‡ Total
4 (22) 4 (22) 4 (22) 2 (11) 2 (11) 1 (6) 1 (6) 18 (100)
*Increase in Paco2 ⬎ 5 mm Hg or decrease in pH ⱖ 0.05. †Any decrease in the Pao2/Fio2 ratio. ‡For heart catheterization and eventual heart surgery. 3920
Complications At least one complication or comorbidity after the start of ventilation occurred in 16 patients (52%) in the NIMV group, vs 23 patients (70%) in the CMV group (relative risk reduction, 26%; 95% CI, ⫺ 3 to 47%; p ⫽ 0.07). Complications were more frequent in the CMV group, but the differences did not reach statistical significance except for the inability to sustain enteral feeding (Table 4). Ventilation Times Geometric means of total ventilation times, adding up the CMV and NIMV times, were 85.9 h and 143.7 h in the NIMV and CMV groups, respectively. The ratio between ventilation times in the NIMV and CMV groups was 0.59 (95% CI, 0.28 to 1.25; p ⫽ 0.17). In patients in whom NIMV failed, the median NIMV time prior to endotracheal intubation was 6.1 h (range, 10 min to 79 h; interquartile range, 2.5 to 24.5 h). Lengths of Stay The geometric means of length of stay in the ICU were 8.9 days (SE 2.6) in the NIMV group and 9.7 days (SE 3.0) in the CMV group. The ratio between the lengths of stay in the NIMV and CMV groups was 0.90 (95% CI, 0.54 to 1.50; p ⫽ 0.69). No statistically significant differences were found in ICU stay or length of stay in hospital when analyzing only those patients who survived (data not shown). Nursing Workload On the first day of ventilation, the TISS-28 score was lower in the NIMV group (median, 27; range, 22 to 60) than the CMV group (median, 32; range, 28 to 46) [p ⫽ 0.007]. The average of the first 3 days was also lower in the NIMV group (median, 27.7; range, 15.3 to 56.3) than the CMV group (median, 31.0; range, 23.3 to 46.7) [p ⫽ 0.03]. The means of direct nursing activity time measurements during the first 8 h were 191 min (SD 66) in the NIMV group and 205 min (SD 63) in the CMV group (p ⫽ 0.48). No differences were found in subsequent measurements. Predictive Factors for the Failure of NIMV In the group that received NIMV, the following parameters considerably improved 60 min after commencing ventilation: Pao2/Fio2 (a mean increase of 32.3, p ⫽ 0.001), respiratory rate (mean decrease of 10.6 breaths/min, p ⫽ 0.001), and the scale of accessory respiratory muscle use (mean decrease, 0.9 points, p ⫽ 0.001). There was a very small decrease Clinical Investigations
Figure 3. Effect of NIMV vs CMV on ICU mortality.
in Paco2 (2.74 mm Hg, p ⫽ 0.24). When these changes in the subgroup that failed were compared with those in whom intubation was avoided, no statistically significant differences were found. The following variables were associated with a higher frequency of endotracheal intubation in the unadjusted analysis: pneumonia as the primary diagnosis (indeterminate odds ratio, p ⫽ 0.001), APACHE II score (odds ratio for every one point increase, 1.20; 95% CI, 1.01 to 1.43; p ⫽ 0.0009), and a respiratory rate ⬍ 35 breaths/min (odds ratio, 7.6; 95% CI, 0.84 to 69.6; p ⫽ 0.038). In the adjusted analysis, pneumonia as the primary diagnosis (odds ratio indeterminate, p ⫽ 0.005) and APACHE II score (odds ratio, 1.20; 95% CI, 0.98 to 1.45; p ⫽ 0.046) jointly predicted the need for endotracheal intubation.
Discussion We found that NIMV avoids intubation in almost 50% of patients with ARF who need mechanical ventilation according to our predefined criteria. This avoidance of intubation might not apply to pneumonia, in which NIMV failed, since all patients were finally intubated. Our results also suggest that NIMV might be associated with lower mortality; however, this could not be confirmed. The extent of therapeutic intervention is lower during the first few days, but nursing workload is equal in the two ventilation modalities. In our patients, NIMV improves oxygenation and reduces respiratory workload measured by the scale of use of accessory muscles and breathing rate. However, Paco2 hardly decreases, a finding also observed by
Figure 4. Patient survival in the ICU depending on the ventilatory modality. www.chestjournal.org
CHEST / 128 / 6 / DECEMBER, 2005
3921
Table 3—Causes of Death Causes
NIMV, No. (%)
CMV, No. (%)
p Value
Septic shock Refractory respiratory failure ARF Arrhythmia Total
3 (30) 5 (50) 1 (10) 1 (10) 10 (100)
7 (50) 6 (43) 0 (0) 1 (7) 14 (100)
0.15 0.62 0.34 0.97 0.30
others.6 This decrease in respiratory workload would be the cause for the improvement,30,35,36 but does not predict the avoidance of intubation since such improvement was alike in the group in which NIMV succeeded and the one in which NIMV failed. Several randomized controlled trials have shown that patients with COPD exacerbations,5,6 acute pulmonary edema,13 hypoxemic ARF,8,10,11,16,37 and ARF of several etiologies7,12 benefit from the addition of NIMV to their usual medical treatment. To date, three metaanalyses comparing NIMV with the usual medical treatment in ARF9,15 and COPD exacerbations9,15,17 have demonstrated benefits from NIMV in reducing endotracheal intubation and mortality, but no benefit was observed in patients with ARF due to causes other than COPD exacerbations. The design of our study differs in two respects. The patients included in our study were more severe, as can be seen by the baseline clinical data, gasometric data, and APACHE II scores; and the aim of our study was to assess whether NIMV could replace CMV in patients with ARF who need it. Accordingly, our patients were severe enough to need mechanical ventilation and had to meet strict criteria so that no patient would be intubated without justification. The choice of our criteria is supported by their high consistency with experienced clinician’s decision.
There are only two previous trials that compared NIMV with CMV and endotracheal intubation in acute respiratory failure. The first trial23 excluded patients with criteria of COPD, and the second trial24 only included patients with acute exacerbations of COPD in whom conventional treatment had failed. Both studies found that NIMV improved gas exchange. The first trial23 found fewer complications and shorter lengths of stay in the ICU in NIMV, whereas the second trial24 did not find any differences in the short-term results. Our trial included patients with ARF without previous COPD and patients with COPD whose causes of ARF were exacerbation and other etiologies. In our stratified analysis, NIMV reduced endotracheal intubation independently of the type of ARF. Mortality in patients in whom NIMV failed was equal to that from patients receiving CMV from the start, which suggests that once strict criteria of NIMV failure are implemented intubation is not dangerously delayed. With regard to nursing workload, our findings of lower therapeutic intervention and similar nursing workload in NIMV compared with CMV support other findings on other tasks in the ICU.38 – 40 Inconsistencies between TISS-28 scores and nursing workloads are more striking in the lowest TISS-28 scores, which despite measuring well the severity and complexity of treatment only explain 43% of nursing direct time.39 The main limitation of our study was its small sample size, which decreased its ability to detect differences in mortality and complications. The study was stopped before achieving the planned sample size due to an ever-decreasing rate of enrollment. Although early termination of a trial is commonly justified on ethical reasons,41 either for a high rate of severe side effects or for a remarking effect of one arm of the trial, there are trials42,43 in ICUs that
Table 4 —Complications Complications
NIMV (n ⫽ 31), No. (%)
CMV (n ⫽ 33), No. (%)
p Value
Shock Sepsis ARF Mental disorders Pneumonia Inability to sustain enteral feeding Facial skin lesions Neuromuscular disorders Cardiac arrest Pneumothorax, pneumomediastinum, subcutaneous emphysema Repeat ventilation (after 48 h) Acute myocardial infarction/angina Laryngeal or tracheal lesions Other complications*
9 (29) 8 (26) 6 (19) 5 (16) 3 (10) 2 (7) 4 (13) 2 (7) 2 (7) 2 (7) 1 (3) 1 (3) 1 (3) 1 (3)
12 (36) 10 (30) 11 (33) 7 (21) 7 (21) 7 (21) 1 (3) 3 (9) 3 (9) 2 (6) 2 (6) 2 (6) 1 (3) 6 (21)
0.42 0.56 0.08 0.72 0.12 0.05 0.15 0.63 0.62 0.90 0.51 0.59 0.97 0.15
*In the NIMV group, pulmonary infection; in the CMV group, sinusitis, arrhythmia, upper digestive hemorrhage, self-extubation, reintubation, repeat ventilation for systemic inflammatory response syndrome, and acute cholecystitis (seven complications in six patients). 3922
Clinical Investigations
were terminated before the planned sample size was achieved because of a decrease in the rate of enrollment. In our study, the known benefit of NIMV when compared with the usual medical treatment in acute exacerbations of COPD may have deterred some physicians from enrolling potential participants since they would expect better results from NIMV than CMV. The external validity of our study might be affected since less severe patients would have not been recruited as much as those more severe. However, as both groups were allocated by chance, internal validity would not be affected. In spite of the fact that our study was a randomized controlled trial and, therefore, the two study groups should have been homogeneous, they were not entirely similar. However, baseline differences between groups did not affect the main results, as can be seen in the stratified analyses. Neither can we rule out the presence of unknown confounders. As our trial was unmasked, the assessment of the need for intubation in NIMV patients could potentially have been biased or delayed despite the predefined criteria for failure. However, we think this was unlikely since in patients from the NIMV group who eventually were intubated, not only were their ventilation times before intubation short but also their mortality was equal to that from patients assigned to CMV (39%). Another consequence of the trial not being masked is that co-intervention might also have happened, but we think it unlikely that any potential co-intervention could have influenced on intubation. The relative effect of NIMV on mortality, clinical complications, and consumption of resources when compared with CMV are issues that need to be confirmed in further trials. Our results might be applied to patients with ARF of several etiologies, but their application to conditions sparsely included in our study, such as ARDS, remains to be seen. Our finding that NIMV in patients with pneumonia, with or without COPD, fails to avoid invasive mechanical ventilation should be looked at with caution since subgroup analyses were not preplanned in the design and this finding was unexpected. Its validity will have to be confirmed in further studies. In conclusion, we found that NIMV reduces the need for intubation in severe ARF with the possible exception of pneumonia. We also found a nonsignificant trend of reduction in ICU and hospital mortality together with fewer complications during ICU stay. The extent of therapeutic intervention is lower during the first few days and does not require longer nursing times.
www.chestjournal.org
Appendix EMVIRA (Ensayo de Modalidades de Ventilacio´n en Insuficiencia Respiratoria Aguda) Investigators Hospital de Mo´stoles: T. Honrubia (principal investigator), N. Franco, M. Mas, P. Galdos, S. Bu´a, J.I. Aria; Hospital Prı´ncipe de Asturias: M. Daguerre, M.A. Mun˜oz, C. Galindo; Fundacio´n Hospital de Alcorco´n: A. Algora, E. Calvo, F. Gordo; Hospital General Yagu¨e: M.J. Lo´pez-Pueyo, L. Cerezo; Hospital Universitario de Getafe: I. Alı´a; Hospital Severo Ochoa: J. Sua´rez J, I. de Miguel, R. Dı´az; Hospital Morales Messeguer: M.D. Rodrı´guez, G. Gonza´lez; Hospital Puerta de Hierro (Clinical Epidemiology): F. Garcı´a Lo´pez, M. Guevara; Data Safety and Monitoring Committee: I. Milla´n, J.J. Rubio, A. Go´mez de la Ca´mara.
References 1 Esteban A, Anzueto A, Frutos F, et al. Characteristics and outcomes in adult patients receiving mechanical ventilation: a 28-day international study. JAMA 2002; 287:345–355 2 Stauffer JI, Olson DE, Petty TL. Complications and consequences of endotracheal intubation and tracheotomy: a prospective study of 150 critically ill adult patients. Am J Med 1980; 70:65–76 3 Pingleton S. Complications of acute respiratory failure. Am J Respir Crit Care Med 1988; 137:1463–1493 4 Fagon J, Chastre J, Hance A. Nosocomial pneumonia in ventilated patients a cohort study evaluating attributable mortality and hospital stay. Am J Med 1993; 94:281–288 5 Bott J, Carroll MP, Conway JH, et al. Randomised controlled trial of nasal ventilation in acute ventilatory failure due to chronic obstructive airways disease. Lancet 1993; 341:1555– 1557 6 Brochard L, Mancebo J, Wysocki M, et al. Noninvasive ventilation for acute exacerbations of chronic obstructive pulmonary disease. N Engl J Med 1995; 333:817– 822 7 Kramer N, Meyer T, Merhag J, et al. Randomized, prospective trial of noninvasive positive pressure ventilation in acute respiratory failure. Am J Respir Crit Care Med 1995; 151: 1799 –1806 8 Wysocki M, Tric L, Wolff MA, et al. Noninvasive pressure support ventilation in patients with acute respiratory failure: a randomized comparison with conventional therapy. Chest 1995; 107:761–768 9 Keenan SP, Kernerman PD, Cook DJ, et al. Effect of noninvasive positive pressure ventilation on mortality in patients admitted with acute respiratory failure: a metaanalysis. Crit Care Med 1997; 25:1685–1692 10 Confalonieri M, Potena A, Carbone G, et al. Acute respiratory failure in patients with severe community-acquired pneumonia: a prospective randomized evaluation of noninvasive ventilation. Am J Respir Crit Care Med 1999; 160:1585–1591 11 Antonelli M, Conti G, Bufi M, et al. Noninvasive ventilation for treatment of acute respiratory failure in patients undergoing solid organ transplantation: a randomized trial. JAMA 2000; 283:235–241 12 Martin TJ, Hovis JD, Costantino JP, et al. A randomized, prospective evaluation of noninvasive ventilation for acute respiratory failure. Am J Respir Crit Care Med 2000; 161: 807– 813 13 Masip J, Betbese´ AJ, Pa´ez J, et al. Non-invasive pressure support ventilation versus conventional oxygen therapy in acute cardiogenic pulmonary oedema: a randomised trial. Lancet 2000; 356:2126 –2132 14 Hilbert G, Gruson D, Vargas F, et al. Noninvasive ventilation CHEST / 128 / 6 / DECEMBER, 2005
3923
15 16 17
18
19 20
21
22
23
24
25 26
27 28
in immunosuppressed patients with pulmonary infiltrates, fever, and acute respiratory failure. N Engl J Med 2001; 344:481– 487 Peter JV, Moran JL, Phillips-Hughes J, et al. Noninvasive ventilation in acute respiratory failure: a meta-analysis update. Crit Care Med 2002; 30:555–562 Ferrer M, Esquinas A, Leo´n M, et al. Noninvasive ventilation in severe hypoxemic respiratory failure: a randomized clinical trial. Am J Respir Crit Care Med 2003; 168:1438 –1444 Lightowler JV, Wedzicha JA, Elliott MW, et al. Non-invasive positive pressure ventilation to treat respiratory failure resulting from exacerbations of chronic obstructive pulmonary disease: Cochrane systematic review and meta-analysis. BMJ 2003; 326:185–187 Ram FS, Lightowler JV, Wedzicha JA. Non-invasive positive pressure ventilation for treatment of respiratory failure due to exacerbations of chronic obstructive pulmonary disease (Cochrane Review). Cochrane Database Syst Rev 2003; CD004104 Wysocki M, Tric L, Wolff MA, et al. Noninvasive pressure support ventilation in patients with acute respiratory failure. Chest 1993; 103:907–913 Antonelli M, Conti G, Riccioni L, et al. Noninvasive positivepressure ventilation via face mask during bronchoscopy with BAL in high-risk hypoxemic patients. Chest 1996; 110:724 – 728 Meduri GU, Turner RE, Abou-Shala N, et al. Noninvasive positive pressure ventilation via face mask: first-line intervention in patients with acute hypercapnic and hypoxemic respiratory failure. Chest 1996; 109:179 –193 Patrick W, Webster K, Ludwig L, et al. Noninvasive positivepressure ventilation in acute respiratory distress without prior chronic respiratory failure. Am J Respir Crit Care Med 1996; 153:1005–1011 Antonelli M, Conti G, Rocco M, et al. A comparison of noninvasive positive-pressure ventilation and conventional mechanical ventilation in patients with acute respiratory failure. N Engl J Med 1998; 339:429 – 435 Conti G, Antonelli M, Navalesi P, et al. Noninvasive vs. conventional mechanical ventilation in patients with chronic obstructive pulmonary disease after failure of medical treatment in the ward: a randomized trial. Intensive Care Med 2002; 28:1701–1707 Girou E, Schortgen F, Delclaux C, et al. Association of noninvasive ventilation with nosocomial infections and survival in critically ill patients. JAMA 2000; 284:2361–2367 Bernard GR, Artigas A, Brigham KL, et al. Report of the American-European Consensus conference on acute respiratory distress syndrome: definitions, mechanisms, relevant outcomes, and clinical trial coordination. Consensus Committee. J Crit Care 1994; 9:72– 81 Mehta S, Jay GD, Woolard RH, et al. Randomized, prospective trial of bilevel versus continuous positive airway pressure in acute pulmonary edema. Crit Care Med 1997; 25:620 – 628 Bartlett JG, Breiman RF, Mandell LA, et al. Community-
3924
29
30
31 32
33 34 35 36
37
38 39 40 41
42
43
acquired pneumonia in adults: guidelines for management. The Infectious Diseases Society of America. Clin Infect Dis 1998; 26:811– 838 Kelly BJ, Matthay MA. Prevalence and severity of neurologic dysfunction in critically ill patients: influence on need for continued mechanical ventilation. Chest 1993; 104:1818 – 1824 Appendini L, Patessio A, Zanaboni S, et al. Physiologic effects of positive end-expiratory pressure and mask pressure support during exacerbations of chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1994; 149:1069 –1076 Bone RC, Sibbald WJ, Sprung CL. The ACCP-SCCM consensus conference on sepsis and organ failure. Chest 1992; 101:1481–1483 Vincent JL, Moreno R, Takala J, et al. The SOFA (Sepsisrelated Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med 1996; 22:707–710 Miranda DR, de Rijk A, Schaufeli W. Simplified Therapeutic Intervention Scoring System: the TISS-28 items–results from a multicenter study. Crit Care Med 1996; 24:64 –73 Francis B, Green M, Payne C. The GLIM System, release 4 manual. Oxford, UK: Clarendon Press, 1984; 531–572 Brochard L. Noninvasive pressure support ventilation: physiological and clinical results in patients with COPD and acute respiratory failure. Monaldi Arch Chest Dis 1997; 52:64 – 67 Girault C, Richard JC, Chevron V, et al. Comparative physiologic effects of noninvasive assist-control and pressure support ventilation in acute hypercapnic respiratory failure. Chest 1997; 111:1639 –1648 Hilbert G, Gruson D, Vargas F, et al. Noninvasive continuous positive airway pressure in neutropenic patients with acute respiratory failure requiring intensive care unit admission. Crit Care Med 2000; 28:3185–3190 Vitacca M, Clini E, Porta R, et al. Preliminary results on nursing workload in a dedicated weaning center. Intensive Care Med 2000; 26:796 –799 Miranda DR, Nap R, de Rijk A, et al. Nursing activities score. Crit Care Med 2003; 31:374 –382 Guccione A, Morena A, Pezzi A, et al. I carichi di lavoro infermeristico. Minerva Anestesiol 2004; 70:411– 416 World Medical Association. World Medical Association declaration of Helsinki: ethical principles for medical research involving human subjects. Available at: www.wma.net/e/policy/b3.htmWorld Medical Association Declaration of Helsinki. Accessed November 20, 2005 Hebert PC, Wells G, Blajchman MA, et al. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. Transfusion Requirements in Critical Care Investigators, Canadian Critical Care Trials Group. N Engl J Med 1999; 340:409 – 417 Gattinoni L, Tognoni G, Pesenti A, et al. Effect of prone positioning on the survival of patients with acute respiratory failure. N Engl J Med 2001; 345:568 –573
Clinical Investigations