Noninvasive positive pressure ventilation for acute respiratory failure patients with chronic obstructive pulmonary disease (COPD): an evidence-based analysis
B R McCurdy
Abstract:Objective: The objective of this evidence-based analysis was to examine the effectiveness, safety, and cost-effectiveness of noninvasive positive pressure ventilation (NPPV) in the following patient populations: patients with acute respiratory failure (ARF) due to acute exacerbations of chronic obstructive pulmonary disease (COPD); weaning of COPD patients from invasive mechanical ventilation (IMV); and prevention of or treatment of recurrent respiratory failure in COPD patients after extubation from IMV. Acute Hypercapnic Respiratory Failure: Respiratory failure occurs when the respiratory system cannot oxygenate the blood and/or remove carbon dioxide from the blood. It can be either acute or chronic and is classified as either hypoxemic (type I) or hypercapnic (type II) respiratory failure. Acute hypercapnic respiratory failure frequently occurs in COPD patients experiencing acute exacerbations of COPD, so this is the focus of this evidence-based analysis. Hypercapnic respiratory failure occurs due to a decrease in the drive to breathe, typically due to increased work to breathe in COPD patients. Technology: There are several treatment options for ARF. Usual medical care (UMC) attempts to facilitate adequate oxygenation and treat the cause of the exacerbation, and typically consists of supplemental oxygen, and a variety of medications such as bronchodilators, corticosteroids, and antibiotics. The failure rate of UMC is high and has been estimated to occur in 10% to 50% of cases. The alternative is mechanical ventilation, either invasive or noninvasive. Invasive mechanical ventilation involves sedating the patient, creating an artificial airway through endotracheal intubation, and attaching the patient to a ventilator. While this provides airway protection and direct access to drain sputum, it can lead to substantial morbidity, including tracheal injuries and ventilator-associated pneumonia (VAP). While both positive and negative pressure noninvasive ventilation exists, noninvasive negative pressure ventilation such as the iron lung is no longer in use in Ontario. Noninvasive positive pressure ventilation provides ventilatory support through a facial or nasal mask and reduces inspiratory work. Noninvasive positive pressure ventilation can often be used intermittently for short periods of time to treat respiratory failure, which allows patients to continue to eat, drink, talk, and participate in their own treatment decisions. In addition, patients do not require sedation, airway defence mechanisms and swallowing functions are maintained, trauma to the trachea and larynx are avoided, and the risk for VAP is reduced. Common complications are damage to facial and nasal skin, higher incidence of gastric distension with aspiration risk, sleeping disorders, and conjunctivitis. In addition, NPPV does not allow direct access to the airway to drain secretions and requires patients to cooperate, and due to potential discomfort, compliance and tolerance may be low. In addition to treating ARF, NPPV can be used to wean patients from IMV through the gradual removal of ventilation support until the patient can breathe spontaneously. Five to 30% of patients have difficultly weaning. Tapering levels of ventilatory support to wean patients from IMV can be achieved using IMV or NPPV. The use of NPPV helps to reduce the risk of VAP by shortening the time the patient is intubated. Following extubation from IMV, ARF may recur, leading to extubation failure and the need for reintubation, which has been associated with increased risk of nosocomial pneumonia and mortality. To avoid these complications, NPPV has been proposed to help prevent ARF recurrence and/or to treat respiratory failure when it recurs, thereby preventing the need for reintubation. Research Questions: What is the effectiveness, cost-effectiveness, and safety of NPPV for the treatment of acute hypercapnic respiratory failure due to acute exacerbations of COPD compared withusual medical care, andinvasive mechanical ventilation?What is the effectiveness, cost-effectiveness, and safety of NPPV compared with IMV in COPD patients after IMV for the following purposes:weaning COPD patients from IMV,preventing ARF in COPD patients after extubation from IMV, andtreating ARF in COPD patients after extubation from IMV? Literature Search: A literature search was performed on December 3, 2010 using OVID MEDLINE, MEDLINE In-Process and Other Non-Indexed Citations, OVID EMBASE, the Cumulative Index to Nursing & Allied Health Literature (CINAHL), Wiley Cochrane, and the Centre for Reviews and Dissemination/International Agency for Health Technology Assessment (INAHTA) for studies published from January 1, 2004 until December 3, 2010. Abstracts were reviewed by a single reviewer and, for those studies meeting the eligibility criteria, full-text articles were obtained. Reference lists were also examined for any additional relevant studies not identified through the search. Since there were numerous studies that examined the effectiveness of NPPV for the treatment of ARF due to exacerbations of COPD published before 2004, pre-2004 trials which met the inclusion/exclusion criteria for this evidence-based review were identified by hand-searching reference lists of included studies and systematic reviews. Inclusion Criteria: English language full-reports;health technology assessments, systematic reviews, meta-analyses, and randomized controlled trials (RCTs);studies performed exclusively in patients with a diagnosis of COPD or studies performed with patients with a mix of conditions if results are reported for COPD patients separately;patient population: (Question 1) patients with acute hypercapnic respiratory failure due to an exacerbation of COPD; (Question 2a) COPD patients being weaned from IMV; (Questions 2b and 2c) COPD patients who have been extubated from IMV. Exclusion Criteria: < 18 years of ageanimal studiesduplicate publicationsgrey literaturestudies examining noninvasive negative pressure ventilationstudies comparing modes of ventilationstudies comparing patient-ventilation interfacesstudies examining outcomes not listed below, such as physiologic effects including heart rate, arterial blood gases, and blood pressure Outcomes of Interest: mortalityintubation rateslength of stay (intensive care unit [ICU] and hospital)health-related quality of lifebreathlessnessduration of mechanical ventilationweaning failurecomplicationsNPPV tolerance and compliance Statistical Methods: When possible, results were pooled using Review Manager 5 Version 5.1, otherwise, the results were summarized descriptively. Dichotomous data were pooled into relative risks using random effects models and continuous data were pooled using weighted mean differences with a random effects model. Analyses using data from RCTs were done using intention-to-treat protocols; P values < 0.05 were considered significant. A priori subgroup analyses were planned for severity of respiratory failure, location of treatment (ICU or hospital ward), and mode of ventilation with additional subgroups as needed based on the literature. Post hoc sample size calculations were performed using STATA 10.1. Quality of Evidence: The quality of each included study was assessed taking into consideration allocation concealment, randomization, blinding, power/sample size, withdrawals/dropouts, and intention-to-treat analyses. The quality of the body of evidence was assessed as high, moderate, low, or very low according to the GRADE Working Group criteria. The following definitions of quality were used in grading the quality of the evidence: [Table: see text] NPPV Plus Usual Medical Care Versus Usual Medical Care Alone for First Line Treatment: A total of 1,000 participants were included in 11 RCTs; the sample size ranged from 23 to 342. The mean age of the participants ranged from approximately 60 to 72 years of age. Based on either the Global Initiative for Chronic Obstructive Lung Disease (GOLD) COPD stage criteria or the mean percent predicted forced expiratory volume in 1 second (FEV1), 4 of the studies included people with severe COPD, and there was inadequate information to classify the remaining 7 studies by COPD severity. The severity of the respiratory failure was classified into 4 categories using the study population mean pH level as follows: mild (pH ≥ 7.35), moderate (7.30 ≤ pH < 7.35), severe (7.25 ≤ pH < 7.30), and very severe (pH < 7.25). Based on these categories, 3 studies included patients with a mild respiratory failure, 3 with moderate respiratory failure, 4 with severe respiratory failure, and 1 with very severe respiratory failure. The studies were conducted either in the ICU (3 of 11 studies) or general or respiratory wards (8 of 11 studies) in hospitals, with patients in the NPPV group receiving bilevel positive airway pressure (BiPAP) ventilatory support, except in 2 studies, which used pressure support ventilation and volume cycled ventilation, respectively. Patients received ventilation through nasal, facial, or oronasal masks. All studies specified a protocol or schedule for NPPV delivery, but this varied substantially across the studies. For example, some studies restricted the amount of ventilation per day (e.g., 6 hours per day) and the number of days it was offered (e.g., maximum of 3 days); whereas, other studies provided patients with ventilation for as long as they could tolerate it and recommended it for much longer periods of time (e.g., 7 to 10 days). These differences are an important source of clinical heterogeneity between the studies. In addition to NPPV, all patients in the NPPV group also received UMC. Usual medical care varied between the studies, but common medications included supplemental oxygen, bronchodilators, corticosteroids, antibiotics, diuretics, and respiratory stimulators. The individual quality of the studies ranged. Common methodological issues included lack of blinding and allocation concealment, and small sample sizes. Need for Endotracheal Intubation Eleven studies reported the need for endotracheal intubation as an outcome. The pooled results showed a significant reduction in the need for endotracheal intubation in the NPPV plus UMC group compared with the UMC alone group (relative risk [RR], 0.38; 95% confidence interval [CI], 0.28−0.50). When subgrouped by severity of respiratory failure, the results remained significant for the mild, severe, and very severe respiratory failure groups. GRADE: moderate Inhospital Mortality Nine studies reported inhospital mortality as an outcome. The pooled results showed a significant reduction in inhospital mortality in the NPPV plus UMC group compared with the UMC group (RR, 0.53; 95% CI, 0.35−0.81). When subgrouped by severity of respiratory failure, the results remained significant for the moderate and severe respiratory failure groups. GRADE: moderate Hospital Length of Stay Eleven studies reported hospital length of stay (LOS) as an outcome. The pooled results showed a significant decrease in the mean length of stay for the NPPV plus UMC group compared with the UMC alone group (weighted mean difference [WMD], −2.68 days; 95% CI, −4.41 to −0.94 days). When subgrouped by severity of respiratory failure, the results remained significant for the mild, severe, and very severe respiratory failure groups. GRADE: moderate Complications Five studies reported complications. Common complications in the NPPV plus UMC group included pneumonia, gastrointestinal disorders or bleeds, skin abrasions, eye irritation, gastric insufflation, and sepsis. Similar complications were observed in the UMC group including pneumonia, sepsis, gastrointestinal disorders or bleeds, pneumothorax, and complicated endotracheal intubations. Many of the more serious complications in both groups occurred in those patients who required endotracheal intubation. Three of the studies compared complications in the NPPV plus UMC and UMC groups. While the data could not be pooled, overall, the NPPV plus UMC group experienced fewer complications than the UMC group. GRADE: low Tolerance/Compliance Eight studies reported patient tolerance or compliance with NPPV as an outcome. NPPV intolerance ranged from 5% to 29%. NPPV tolerance was generally higher for patients with more severe respiratory failure. Compliance with the NPPV protocol was reported by 2 studies, which showed compliance decreases over time, even over short periods such as 3 days. NPPV Versus IMV for the Treatment of Patients Who Failed Usual Medical Care: A total of 205 participants were included in 2 studies; the sample sizes of these studies were 49 and 156. The mean age of the patients was 71 to 73 years of age in 1 study, and the median age was 54 to 58 years of age in the second study. Based on either the GOLD COPD stage criteria or the mean percent predicted FEV1, patients in 1 study had very severe COPD. The COPD severity could not be classified in the second study. Both studies had study populations with a mean pH less than 7.23, which was classified as very severe respiratory failure in this analysis. One study enrolled patients with ARF due to acute exacerbations of COPD who had failed medical therapy. The patient population was not clearly defined in the second study, and it was not clear whether they had to have failed medical therapy before entry into the study. Both studies were conducted in the ICU. Patients in the NPPV group received BiPAP ventilatory support through nasal or full facial masks. Patients in the IMV group received pressure support ventilation. Common methodological issues included small sample size, lack of blinding, and unclear methods of randomization and allocation concealment. Due to the uncertainty about whether both studies included the same patient population and substantial differences in the direction and significance of the results, the results of the studies were not pooled. Mortality Both studies reported ICU mortality. Neither study showed a significant difference in ICU mortality between the NPPV and IMV groups, but 1 study showed a higher mortality rate in the NPPV group (21.7% vs. 11.5%) while the other study showed a lower mortality rate in the NPPV group (5.1% vs. 6.4%). One study reported 1-year mortality and showed a nonsignificant reduction in mortality in the NPPV group compared with the IMV group (26.1% vs. 46.1%). GRADE: low to very low Intensive Care Unit Length of Stay Both studies reported LOS in the ICU. The results were inconsistent. One study showed a statistically significant shorter LOS in the NPPV group compared with the IMV group (5 ± 1.35 days vs. 9.29 ± 3 days; P < 0.001); whereas, the other study showed a nonsignificantly longer LOS in the NPPV group compared with the IMV group (22 ± 19 days vs. 21 ± 20 days; P = 0.86). GRADE: very low Duration of Mechanical Ventilation Both studies reported the duration of mechanical ventilation (including both invasive and noninvasive ventilation). The results were inconsistent. One study showed a statistically significant shorter duration of mechanical ventilation in the NPPV group compared with the IMV group (3.92 ± 1.08 days vs. 7.17 ± 2.22 days; P < 0.001); whereas, the other study showed a nonsignificantly longer duration of mechanical ventilation in the NPPV group compared with the IMV group (16 ± 19 days vs. 15 ± 21 days; P = 0.86). GRADE: very low Complications Both studies reported ventilator-associated pneumonia and tracheotomies. Both showed a reduction in ventilator-associated pneumonia in the NPPV group compared with the IMV group, but the results were only significant in 1 study (13% vs. 34.6%, P = 0.07; and 6.4% vs. 37.2%, P < 0.001, respectively). Similarly, both studies showed a reduction in tracheotomies in the NPPV group compared with the IMV group, but the results were only significant in 1 study (13% vs. 23.1%, P = 0.29; and 6.4% vs. 34.6%; P < 0.001). GRADE: very low Other Outcomes One of the studies followed patients for 12 months. At the end of follow-up, patients in the NPPV group had a significantly lower rate of needing de novo oxygen supplementation at home. In addition, the IMV group experienced significant increases in functional limitations due to COPD, while no increase was seen in the NPPV group. Finally, no significant differences were observed for hospital readmissions, ICU readmissions, and patients with an open tracheotomy, between the NPPV and IMV groups. NPPV for Weaning COPD Patients From IMV: A total of 80 participants were included in the 2 RCTs; the sample sizes of the studies were 30 and 50 patients. The mean age of the participants ranged from 58 to 69 years of age. Based on either the GOLD COPD stage criteria or the mean percent predicted FEV1, both studies included patients with very severe COPD. Both studies also included patients with very severe respiratory failure (mean pH of the study populations was less than 7.23). Chronic obstructive pulmonary disease patients receiving IMV were enrolled in the study if they failed a T-piece weaning trial (spontaneous breathing test), so they could not be directly extubated from IMV. Both studies were conducted in the ICU. Patients in the NPPV group received weaning using either BiPAP or pressure support ventilation NPPV through a face mask, and patients in the IMV weaning group received pressure support ventilation. In both cases, weaning was achieved by tapering the ventilation level. The individual quality of the studies ranged. Common methodological problems included unclear randomization methods and allocation concealment, lack of blinding, and small sample size. Mortality Both studies reported mortality as an outcome. The pooled results showed a significant reduction in ICU mortality in the NPPV group compared with the IMV group (RR, 0.47; 95% CI, 0.23−0.97; P = 0.04). GRADE: moderate Intensive Care Unit Length of Stay Both studies reported ICU LOS as an outcome. The pooled results showed a nonsignificant reduction in ICU LOS in the NPPV group compared with the IMV group (WMD, −5.21 days; 95% CI, −11.60 to 1.18 days). GRADE: low Duration of Mechanical Ventilation Both studies reported duration of mechanical ventilation (including both invasive and noninvasive ventilation) as an outcome. The pooled results showed a nonsignificant reduction in duration of mechanical ventilation (WMD, −3.55 days; 95% CI, −8.55 to 1.44 days). GRADE: low Nosocomial Pneumonia Both studies reported nosocominal pneumonia as an outcome. The pooled results showed a significant reduction in nosocomial pneumonia in the NPPV group compared with the IMV group (RR, 0.14; 95% CI, 0.03−0.71; P = 0.02). GRADE: moderate Weaning Failure One study reported a significant reduction in weaning failure in the NPPV group compared with the IMV group, but the results were not reported in the publication. In this study, 1 of 25 patients in the NPPV group and 2 of 25 patients in the IMV group could not be weaned after 60 days in the ICU. NPPV After Extubation of COPD Patients From IMV: The literature was reviewed to identify studies examining the effectiveness of NPPV compared with UMC in preventing recurrence of ARF after extubation from IMV or treating acute ARF which has recurred after extubation from IMV. No studies that included only COPD patients or reported results for COPD patients separately were identified for the prevention of ARF postextubation. One study was identified for the treatment of ARF in COPD patients that recurred within 48 hours of extubation from IMV. This study included 221 patients, of whom 23 had COPD. A post hoc subgroup analysis was conducted examining the rate of reintubation in the COPD patients only. A nonsignificant reduction in the rate of reintubation was observed in the NPPV group compared with the UMC group (7 of 14 patients vs. 6 of 9 patients, P = 0.67). GRADE: low NPPV Plus UMC Versus UMC Alone for First Line Treatment of ARF due to Acute Exacerbations of COPD: Moderate quality of evidence showed that compared with UMC, NPPV plus UMC significantly reduced the need for endotracheal intubation, inhospital mortality, and the mean length of hospital stay.Low quality of evidence showed a lower rate of complications in the NPPV plus UMC group compared with the UMC group. NPPV Versus IMV for the Treatment of ARF in Patients Who Have Failed UMC: Due to inconsistent and low to very low quality of evidence, there was insufficient evidence to draw conclusions on the comparison of NPPV versus IMV for patients who failed UMC. NPPV for Weaning COPD Patients From IMV: Moderate quality of evidence showed that weaning COPD patients from IMV using NPPV results in significant reductions in mortality, nosocomial pneumonia, and weaning failure compared with weaning with IMV.Low quality of evidence showed a nonsignificant reduction in the mean LOS and mean duration of mechanical ventilation in the NPPV group compared with the IMV group. NPPV for the Treatment of ARF in COPD Patients After Extubation From IMV: Low quality of evidence showed a nonsignificant reduction in the rate of reintubation in the NPPV group compared with the UMC group; however, there was inadequate evidence to draw conclusions on the effectiveness of NPPV for the treatment of ARF in COPD patients after extubation from IMV