Criteria abstracted from The Users' Guide to Medical Literature, from the Health Information Research Unit and Clinical Epidemiology and Biostatistics, McMaster University

Highlighted lines and questions below provide links to the pertinent description of criteria in The EBM User's Guide, now available at the Canadian Centres for Health Evidence

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

Does noninvasive positive pressure ventilation improve outcome in acute hypoxemic respiratory failure? A systematic review.

Keenan SP, Sinuff T, Cook DJ, Hill NS.

Crit Care Med. 2004 Dec;32(12):2516-23. [abstract]

Reviewed by Karl Yen MD, Children's Hospital, Boston

Review posted April 13, 2005

---

I. Are the results of the study valid?

A. Primary questions:

1. Did the overview address a focused clinical question?

Yes. The question being addressed is: does noninvasive positive pressure ventilation (NPPV) improve outcome in acute hypoxemic failure?

Specifically, the study's stated objective is "to systemically review the RCTs of patients with acute hypoxemic respiratory failure unrelated to cardiogenic pulmonary edema, to determine the effect of the addition of NPPV to standard therapy on endotracheal intubation, intensive care unit (ICU) and hospital length of stay, and mortality."

2. Were the criteria used to select articles for inclusion appropriate?

Yes. The articles were selected using criteria well suited to answering the clinical question. The selection criteria included: a) randomized controlled trials; b) study populations comprised mostly of patients with acute hypoxemic respiratory failure not associated with cardiogenic pulmonary edema and not associated with chronic obstructive pulmonary disease (COPD) exacerbation and also not requiring immediate ventilatory support; c) comparison of noninvasive ventilation and standard therapy vs. standard therapy alone; d) clinical outcomes included the need for endotracheal intubation, length of ICU or hospital stay, or ICU or hospital mortality.

B. Secondary questions:

3. Is it unlikely that important, relevant studies were missed?

Yes. The authors searched the major databases, including MEDLINE, EMBASE, and the Cochrane Library, limiting their search to RCTs using relevant keywords and without language restriction. They hand-searched the abstracts from key meetings and reviewed the reference lists of all identified studies and reviews as well as their personal files. Furthermore, the authors contacted experts and first authors of selected articles to help identify published and unpublished studies not identified in their search strategies. Funnel plots were created to estimate the potential influence of unpublished trials on the results of the systematic review. No publication bias was suggested from this analysis.

4. Was the validity of the included studies appraised?

Yes. Following abstraction of the data, the authors critically appraised the trials using 11 validity criteria. These criteria included the following: randomization, concealment, blinding, patient selection, comparability of groups at baseline, treatment protocol, confounders, cointerventions, outcome definitions, extent of follow-up, and intention-treat analysis. The methodological quality of each of 8 included trials was explicitly presented in a table. Methodological quality scores ranged from 4 to 10 of a possible 11. Given the nature of NPPV as an intervention, none of the studies were blinded. Many of the studies attempted to reduce bias through the use of concealed randomization, cointervention documentation, as well as specific intubation criteria and outcome assessment.

5. Were assessments of studies reproducible?

Yes. Data abstraction was performed by two of the authors independently and in duplicate. They resolved differences in opinion by consensus or by consulting a third investigator.

6. Were the results similar from study to study?

No. Using a p-value cut-off < 0.1, the chi-square test for heterogeneity demonstrated significant heterogeneity between the different study results for the effect of NPPV on endotracheal intubation for patients with acute hypoxemic respiratory failure. This was true for both overall intubation rates (p = 0.09) as well as intubation rates excluding patients with cardiogenic pulmonary edema and/or COPD (p = 0.05) Furthermore, visual inspection of the Forest plots of the study results reveal qualitative differences among study results. These differences were attributed to the diversity of patient population. There were three trials of relatively heterogeneous patients, two trials with immunocompromised patients, and individual trials on post-lung resection surgery patients, community-acquired pneumonia, and post-extubation respiratory failure.

II. What are the results?

1. What are the overall results of the review?

From an initial electronic search of 763 studies, eight fully published randomized controlled trials, with a total of 366 patients, met the authors' selection criteria. The studies were all in English; there were two multi-center and six single-center trials. From the six trials that included patients with COPD or cardiogenic pulmonary edema, the authors were able to obtain patient-specific data on patients without these diagnoses. All studies utilized pressure-cycled ventilation and NPPV was administered via a nasal mask, full-mask, or a combination of the two. Data on the need for endotracheal intubation and ICU mortality was obtained from all eight trials, while other outcomes were not universally measured.

The authors performed outcome assessment on the whole group of patients followed by subgroup analysis of each of the outcomes excluding those patients with COPD and cardiogenic pulmonary edema. Overall, NPPV was associated with a significantly lower rate of endotracheal intubation than standard management, a reduction in length of ICU stay, an increase in length of hospital stay, and a reduction in ICU mortality. There was also a trend towards lower hospital mortality (Table 1). ICU and Hospital mortality rates for the control groups were separately plotted against the risk difference between the intervention and control groups. Although not statistically-based, the post hoc exercise suggests that NPPV improves survival particularly in patients that are known to have a high mortality rate if required to undergo conventional mechanical ventilation such as immunocompromised or post-thoracotomy patients.

Table 1. Effect of NPPV on Clinical Outcomes

	Effect of NPPV on Patients with Acute Hypoxemic Respiratory Failure		Subgroup Analysis Excluding COPD and CPE Patients
Endotracheal Intubation	ARR 23%	95% CI 10-35%	ARR 24%	95% CI 8-36%
Length of ICU Stay	Reduction of 1.9 days	95% CI 1.0-2.9 days	Reduction of 1.9 days	95% CI 1.0-2.9 days
Length of Hospital Stay			Increase of 2.8 days	95% CI 0.9-4.7 days
ICU Mortality	ARR of 17%	95% CI 8-26%	ARR of 16%	95% CI 5-27%
Hospital Mortality			ARR 10%	95% CI (-7)-27%

ARR = Absolute Risk Reduction, CI = Confidence Interval, CPE = cardiogenic pulmonary edema

2. How precise were the results?

The 95% CI was provided for each of the effects of NPPV on clinical outcome. Overall, the results were precise and the lower limit of the risk differences would still be considered clinically significant. In other words, if the ARR for NPPV in these patients was "only" 8%, that would still be a meaningful risk reduction.

III. Will the results help me in caring for my patients?

1. Can the results be applied to my patient care?

No. Since all the studies included in the review were adult-based, the results cannot be applied directly to the pediatric intensive care population. At best, we could conceivably extrapolate from the non-CHF/COPD group to our adolescent population with acute hypoxemic respiratory failure. Variable inter-center NPPV experience would be expected to impact implementation as well as results.

2. Were all clinically important outcomes considered?

No. Clinically important outcomes including mortality, endotracheal intubation, and length of stay were considered. However, the study did not address other clinically important outcomes such as cost, complications, patient comfort, critical care myopathy, sedation requirements, and nutrition.

3. Are the benefits worth the harms and costs?

A determination of overall benefit cannot be made based on this study. Although the results of this systematic review suggest that NPPV is beneficial in patients with acute hypoxemic respiratory failure, conclusions are difficult to draw given limitations such as study size and heterogeneity. Specific populations such as immunocompromised patients may derive particular benefit from the application of NPPV. Large scale trials of NPPV in such populations may be warranted. On the other hand, subgroups of patients with acute hypoxemic respiratory failure not analyzed in this systemic review may be harmed by NPPV. Patients developing respiratory distress post-extubation outside the setting of lung resection, in particular, may not benefit from NPPV as shown in a RCT performed by Keenan et al (1). Finally, as noted in the response to the preceding question, the systematic review failed to consider several potential harms as well as cost.

References:

1. Keenan SP, Powers C, McCormack DG, et al. Nonininvasive positive-pressure ventilation for postextubation respiratory distress: a randomized controlled trial. JAMA 2002; 287:3238-3244 [abstract]

Comments

Back to the EB Journal Club Index