Saladeaula.abitep.com.br

Noninvasive Ventilation in Acute Cardiogenic
Pulmonary Edema: Systematic Review and
Online article and related content Josep Masip; Marta Roque; Bernat Sánchez; et al. current as of August 27, 2008. JAMA. 2005;294(24):3124-3130 (doi:10.1001/jama.294.24.3124) Topic collections Quality of Care; Evidence-Based Medicine; Review; Critical Care/ Intensive CareMedicine; Adult Critical Care Noninvasive Ventilation in
Acute Cardiogenic Pulmonary Edema
Systematic Review and Meta-analysis
Context In patients with acute cardiogenic pulmonary edema noninvasive ventila-
tion may reduce intubation rate, but the impact on mortality and the superiority of Bernat Sa´nchez, MD one technique over another have not been clearly established.
Rafael Ferna´ndez, MD Objective To systematically review and quantitatively synthesize the short-term effect
of noninvasive ventilation on major clinical outcomes.
Mireia Subirana, RN Data Sources MEDLINE and EMBASE (from inception to October 2005) and Coch-
Jose´ Angel Expo´sito, BSc rane databases (library issue 4, 2005) were searched to identify relevant randomized controlled trials and systematic reviews published from January 1, 1988, to October ONINVASIVE VENTILATION (NIV) is a modality of venti- Study Selection and Data Extraction Included trials were all parallel studies com-
latory support without endo- paring noninvasive ventilation to conventional oxygen therapy in patients with acute pul- tracheal intubation and seda- monary edema. Comparisons of different techniques, either continuous positive airway tion that has demonstrated to be useful pressure (CPAP) or bilevel noninvasive pressure support ventilation (NIPSV), were also in several forms of respiratory failure. In patients with severe exacerbation of Data Synthesis Fifteen trials were selected. Overall, noninvasive ventilation signifi-
chronic obstructive pulmonary disease, cantly reduced the mortality rate by nearly 45% compared with conventional therapy it has been shown to reduce mortal- (risk ratio [RR], 0.55; 95% confidence interval [CI], 0.40-0.78; P=.72 for heterogene- ity.1,2 In the setting of acute pulmonary ity). The results were significant for CPAP (RR, 0.53; 95% CI, 0.35-0.81; P=.44 for het- edema, NIV has been shown to reduce erogeneity) but not for NIPSV (RR, 0.60; 95% CI, 0.34-1.05; P=.76 for heterogeneity), the intubation rate in several random- although there were fewer studies in the latter. Both modalities showed a significant ized trials, either using continuous posi- decrease in the "need to intubate" rate compared with conventional therapy: CPAP (RR,0.40; 95% CI, 0.27-0.58; P=.21 for heterogeneity), NIPSV (RR, 0.48; 95% CI, 0.30- tive airway pressure (CPAP)3-8 or bi- 0.76; P=.24 for heterogeneity), and together (RR, 0.43; 95% CI, 0.32-0.57; P=.20 for level noninvasive pressure support heterogeneity). There were no differences in intubation or mortality rates in the analysis ventilation (NIPSV).8,9 The technique of of studies comparing the 2 techniques.
CPAP is simpler and may be performed Conclusions Noninvasive ventilation reduces the need for intubation and mortality
with an oxygen source connected to a in patients with acute cardiogenic pulmonary edema. Although the level of evidence tight-fitting face mask or helmet, with an is higher for CPAP, there are no significant differences in clinical outcomes when com- expiratory valve to maintain constant paring CPAP vs NIPSV.
positive intrathoracic pressure. Con- versely, NIPSV is more complex, re-quires a ventilator to provide 2 levels of tice,10,11 but its use is often based ity, and considerable controversy pressure: one to assist patients with in- more on perceived efficacy than on remains over which technique is spiratory positive airway pressure (IPAP) scientific evidence.11 This may be superior to the other.12,13 and the other, like CPAP, to maintain ex- explained because no single trial has We undertook a systematic review piratory positive pressure (EPAP).
shown an impact in hospital mortal- to investigate the effect of NIV on the With widespread adoption of NIV in patients with acute and chronic Author Affiliations: ICU Department, Hospital Dos de
Medicina Intensiva, Hospital de Sabadell, Corporacio´ respiratory failure over the last 2 Maig Consorci Sanitari Integral, University of Barce- Parc Taulı´, Sabadell (Dr Ferna´ndez), Spain.
decades, acute pulmonary edema is lona (Drs Masip and Sa´nchez) and Iberoamerican Coch- Corresponding Author: Josep Masip, MD, ICU De-
currently the second most common rane Center, Hospital de la Santa Creu i Sant Pau, Uni- partment, Hospital Dos de Maig Consorci Sanitari In- versitat Auto´noma de Barcelona (Mss Roque and tegral Barcelona, University of Barcelona, Dos de Maig indication for NIV in clinical prac- Subirana and Mr Expo´sito), Barcelona, and Servei de 301, 08025 Barcelona, Spain ([email protected]).
3124 JAMA, December 28, 2005—Vol 294, No. 24 (Reprinted)
2005 American Medical Association. All rights reserved.
NONINVASIVE VENTILATION IN ACUTE CARDIOGENIC PULMONARY EDEMA main outcomes (intubation and mor-tality) comparing the 2 techniques to Figure 1. Flow Diagram of Trial Selection
each other and to conventional oxy- 559 Trials Identified as Potentially Relevant gen therapy.
and Screened for Retrieval 39 CENTRAL Database 199 EMBASE Database321 MEDLINE Database 532 Trials Excluded as Not Relevant, Not Randomized We aimed to identify all randomized con- Controlled Trial, or Duplicated trolled trials assessing the efficacy of NIV 27 Retrieved for More Detailed Evaluation in patients with acute pulmonary edema.
11 Trials Excluded 1 Design Containing Inadequate Adjusted Planned The electronic search strategy applied standard filters for identification of ran- 2 Crossover Design18,191 Out-of-Hospital Setting With Inappropriate Allocation20 domized clinical trials. Databases 2 Duplicated Data6,23 searched were the Cochrane Central Reg- 1 Subgroup of Patients With Acute Respiratory Failure213 Abstract Proceeding, Unpublished as a Full Paper24-26 ister of Controlled Trials (CENTRAL, 1 Published in Chinese27 The Cochrane Library Issue 4, 2005), 16 Potentially Appropriate Trials to Be MEDLINE (from inception to October Included in the Meta-analysis 2005), and EMBASE (from inception to 1 Trial Excluded Because No Available Data for Outcomes of Interest28 October 2005). We did not apply lan-guage restrictions. In addition to the elec- 15 Trials Included in the tronic search, we checked out cross- 15 With Outcome Data on Need to Intubate15 With Hospital Mortality Data references from original articles and 14 With Myocardial Infarction Data reviews and sometimes contactedauthors to obtain additional unpub-lished data. Our search included the fol- respiratory failure were excluded, as The primary outcomes for the in- lowing: continuous positive airway pres- were studies published only in ab- cluded trials were treatment failure, en- sure (Medical Subject Headings [MeSH]); stract form and those written in a non- dotracheal intubation, myocardial in- continuous positive airway*; biphasic accessible language after failure to ob- farction, resolution time, therapeutical intermittent positive airway; bilevel posi- tain more complete data.
success at 2 hours, 48-hour mortality, in- tive airway*; noninvasive ventilatory- hospital mortality, and specific labora- assistance apparatus; noninvasive sup- tory or physiological parameters. The pri- port ventilation; noninvasive ventilat*; The initial selection was performed by mary outcomes for the present study non-invasive ventilat*; CPAP; Bipap; pul- distributing references among pairs of were treatment failure and in-hospital monary edema (MeSH); acute pulmo- independent reviewers. A full-text copy mortality because all the included trials nary edema; heart failure, congestive of all studies of possible relevance was presented data about these items. How- (MeSH); edema, cardiac (MeSH); acute obtained and data from each study was ever, treatment failure was often re- cardiogenic edema linked with random- extracted independently by paired re- ported using different definitions. It was ized controlled trial OR controlled clini- viewers, using a prestandardized data endotracheal intubation in some stud- cal trial OR randomized controlled trials abstraction form. Data extracted were ies, "criteria for intubation" (which was OR random allocation OR double-blind checked by a third reviewer ( J.M. or not necessarily performed) in others, and method OR single-blind method OR clini- M.R.) for accuracy. The reviewers de- some arbitrary clinical or blood gas cri- cal trial OR clinical trials in various com- cided which trials fitted the inclusion teria at different intervals of time in oth- criteria focusing on study design, pa- ers. For this item, we finally decided to tients' characteristics, protocol of the in- select the variable "need to intubate," Selection of Studies
terventions, outcomes measured, and which included those patients who were We restricted the analysis to parallel main results. Any disagreement appear- intubated and those who needed to be randomized trials comparing NIV to ing during the process was solved by intubated but were not, either due to suc- conventional oxygen therapy or to an- discussion and team consensus.
cessful rescue NIV, patient's refusal, or other NIV modality. Study designs con- Methodological quality of the in- a medical decision on account of seri- taining inadequately adjusted planned cluded trials was assessed collecting ous comorbidities.
cointerventions and crossover trials data on key domains related to valid- Myocardial infarction was consid- were not included. Studies that ana- ity14,15: reporting of allocation conceal- ered a secondary outcome in the pres- lyzed the application of NIV in pa- ment, description of an adequate ran- ent study. This complication was com- tients with acute pulmonary edema as domization method, and specification puted whether it was identified as the a part of a group of patients with acute of loss of subjects.
cause of acute pulmonary edema or was 2005 American Medical Association. All rights reserved.
(Reprinted) JAMA, December 28, 2005—Vol 294, No. 24 3125
NONINVASIVE VENTILATION IN ACUTE CARDIOGENIC PULMONARY EDEMA diagnosed soon after admission. Pre- poxemia was required for diagnosis, preparing and maintaining Cochrane sys- vious episodes of myocardial infarc- whether assessed by pulse-oximetry or tematic reviews. Although the main tion were not counted. Other terms like arterial blood gas samples.
analysis was made considering avail- intensive care unit (ICU) length of stay, able data as finally published by au- hospital length of stay, one-year mor- thors, an intention-to-treat sensitivity tality, physiological measurements at We summarized available data for all analysis was also performed in order to baseline and at 1 hour, and adverse ef- trials reporting results on need to intu- obtain more exact results, assuming that fects were also collected but were not bate or mortality, computing pooled risk lost or withdrawn patients experienced analyzed because there was a lack of this ratios (RRs) and their respective 95% outcomes (either need to intubate or information in many of the studies.
confidence intervals (CIs) by means of death). Three-arm trials were analyzed Although there were heterogeneities a fixed-effects meta-analysis model. We as 2-arm separate trials in each compari- in the definition of acute pulmonary examined heterogeneity using a ␹2 test.
son, duplicating the control group data.
edema, it was generally described as dys- All statistical analyses were performed A sensitivity analysis was performed cor- pnea of acute onset, with physical and with Review Manager (Revman version recting for this artificial sample size in- radiological signs of pulmonary edema.
4.2 for Windows, Oxford, England), the crease, showing no relevant differences In addition, in almost all the studies hy- Cochrane Collaboration's software for with respect to the main analysis.
Table. Randomized Studies Analyzing Noninvasive Ventilation
Continuous Positive Airway Pressure vs Oxygen Therapy
Clinical outcomes 1 ICU in Australia 40 (39) Full face In-hospital mortality Takeda et al,29 1 ICU in Japan 30 (29) Full face or Laboratory parameters Measurement of plasma 1 ED and ICU in the Clinical outcomes Measurement of plasma Laboratory parameters Elderly patients (⬎75 y) Noninvasive Pressure Support Ventilation vs Conventional Oxygen Therapy
40 (37) Full face IPAP was adjusted to tidal 1 ED in the United Prematurely interrupted when the study by Mehtaet al35 was published 14.5/6.1, Mean Intubation Post hoc analysis in hypercapnic patients Trials With 3 Study Groups
Full-face mask for CPAP and 2 EDs in the United Success in ED (2 h) Prehospital nitrates therapy In-hospital mortality 83 (80) Full face Continuous Positive Airway Pressure vs Noninvasive Pressure Support Ventilation
1 ED in the United Prematurely stopped for higher rate of AMI in Bellone et al,36 1 ED in Italy Study restricted to patients Bellone et al,37 1 ED in Italy Primary end point was AMI Only nonischemic APE Abbreviations: AMI, acute myocardial infarction; APE, acute pulmonary edema; CPAP, continuous positive airway pressure; ED, emergency department; EPAP, positive expiratory airway pressure (equivalent to CPAP); ICU, intensive care unit; IPAP, inspiratory positive airway pressure; NIPSV, bilevel noninvasive pressure support ventilation.
*Numbers in parentheses denote the number of patients finally included after withdrawals.
3126 JAMA, December 28, 2005—Vol 294, No. 24 (Reprinted)
2005 American Medical Association. All rights reserved.
NONINVASIVE VENTILATION IN ACUTE CARDIOGENIC PULMONARY EDEMA Publication bias was assessed apply- Figure 2. Effects of Noninvasive Ventilation on Death
ing the Egger et al16 and Begg et al17 sta-tistical tests to the 2 main outcomes of the included trials: intubation and in- No. of Events/ Total No.
hospital mortality. Publication bias was Noninvasive Control studied separately for trials compar- Continuous Positive Airway Pressure ing NIV with control and for trials com- Räsänen et al,3 1985 paring modalities of NIV.
Bersten et al,4 1991 Takeda et al,29 1997 Park et al,30 2001 Kelly et al,31 2002 Our initial electronic search identified Crane et al,32 2004 L‘Her et al,7 2004 559 studies. Of these, 532 were ex- Park et al,8 2004 cluded because they were not random- Risk Ratio, 0.5395% Confidence Interval, 0.35-0.81 ized trials, did not evaluate NIV in pa- tients with acute pulmonary edema, P = .44 for Heterogeneity Noninvasive Pressure Support Ventilation were duplicated references, or were not relevant. Twenty-seven studies were re- Masip et al,9 2000 trieved for more detailed analysis, 11 Park et al,30 2001 of which were excluded. Two were ex- Nava et al,34 2003 Crane et al,32 2004 cluded because of crossover de- Park et al,8 2004 sign18,19; 1 for out-of-hospital setting with inappropriate allocation20; 1 for re- 95% Confidence Interval, 0.34-1.05P = .07 cruitment of patients with acute pul- P = .76 for Heterogeneity monary edema as a part of a series with acute respiratory failure21; 1 for study 95% Confidence Interval, 0.40-0.78P <.001 design containing inadequately ad- P = .72 for Heterogeneity justed planned cointerventions22; 2 for duplicated publications, partial6 or com- Risk Ratio (95% Confidence Interval) plete23; 3 for results reported exclu-sively in proceedings,24-26 and 1 study Data markers are proportional to the amount of data contributed by each trial.
published in a nonaccessible language.27 The flow diagram of the trial selec- tory tract infection, arrhythmia, vol- age IPAP ranged from 14.5 to 20 cm H2O tion process is shown in FIGURE 1. Six-
ume overload, or treatment noncom- with 15 cm H2O being the most re- teen studies were selected, one of which pliance, accounted for 28% of the cases.
peated value. Conversely, EPAP was set was finally excluded because the re- Causes of death were reported in few at 5 cm H2O in most trials. Ventilators ported outcomes did not meet our se- studies5,8,9,30,35,36 and half of the cases used for NIPSV differed substantially lection criteria.28 Thus, we included 15 were related to shock.
from one study to another. Intensive care trials in the meta-analysis.3-5,7-9,29-37 All trials used full face masks (oro- unit ventilators were used in one trial,9 nasal) but nasal masks were also used whereas specific NIV portable ventila- in 27% of them. Nine studies com- tors were used in the others. Early stud- Trial characteristics are summarized in pared CPAP with conventional oxygen ies used very simple devices.
the TABLE. Although all were pub-
therapy,3-5,7,8,29-32 3 of them involved in In general, methodological quality lished in English, they represent an in- a 3-branch design concomitantly ana- was acceptable. Eleven out of 15 trial ternational experience, including data lyzing NIPSV.8,30,32 Six studies com- reports described the use of appropri- from 11 countries. Three studies were pared NIPSV with conventional oxy- ate randomization methods, mainly multiple-center trials, whereas the oth- gen therapy,8,9,30,32-34 3 being those ers were conducted in a single center.
mentioned with 3 branches. Finally, lists.4,7-9,31-37 Nine of the studies de- Causes of acute pulmonary edema 6 studies compared CPAP with scribed the use of a concealed alloca- were reported in 11 of the stud- NIPSV8,30,32,35-37 and again, 3 of them also tion method, all but one using sealed ies3-5,7-9,29,30,34,36,37 and were described as compared conventional therapy. The envelopes with or without external ran- acute coronary syndrome in 203 (31%) CPAP level used in these trials ranged domization,3,7-9,31,32,34,36,37 and all stud- of the patients, hypertension in 178 from 2.5 to 12.5 cm H2O although the ies reported the number of patients, if (27%), or worsening heart failure in 92 most frequent pressure was 10 cm H2O.
any, lost to follow-up. Nine studies in- (14%). Other precipitants like respira- The level of NIPSV was variable. Aver- cluded sample-size calculations.4,7-9,33-37 2005 American Medical Association. All rights reserved.
(Reprinted) JAMA, December 28, 2005—Vol 294, No. 24 3127
NONINVASIVE VENTILATION IN ACUTE CARDIOGENIC PULMONARY EDEMA heterogeneity), and 49% for NIPSV (RR, Figure 3. Effects of Noninvasive Ventilation on Need to Intubate
0.51; 95% CI, 0.33-0.78; P = .42 for Need to Intubate, No. of Events/Total No.
The overall myocardial infarction rate Noninvasive Control for NIV was 78 (22.5%) of 346, which was similar to that observed for conven- Continuous Positive Airway Pressure Räsänen et al,3 1985 tional therapy, 78 (26.8%) of 292 (RR, Bersten et al,4 1991 0.89; 95% CI, 0.69-1.17; P=.99 for het- erogeneity). In about 60% of the cases, Takeda et al,29 1997 Park et al,30 2001 myocardial infarction was reported as a Kelly et al,31 2002 cause of acute pulmonary edema. Ad- Crane et al,32 2004 verse effects like vomiting, abdominal L‘Her et al,7 2004 Park et al,8 2004 distention, claustrophobia, or skin re- actions were infrequent and were re- 95% Confidence Interval, 0.27-0.58P <.001 ported only in a few patients.
P = .21 for Heterogeneity Comparison Between CPAP and
Noninvasive Pressure Support Ventilation NIPSV. Six studies compared CPAP with
Masip et al,9 2000 NIPSV, and 3 of these also compared the Park et al,30 2001 2 techniques with conventional treat- Nava et al,34 2003 ment. Overall, the number of patients Crane et al,32 2004 included in these studies was 219. No Park et al,8 2004 differences were seen in the main out- 95% Confidence Interval, 0.30-0.76 comes, mortality and need-to-intubate P = .002P = .24 for Heterogeneity rate, in the studies comparing CPAP to NIPSV (FIGURE 4). Although a slight ten-
95% Confidence Interval, 0.32-0.57 dency in favor of NIPSV was observed P <.001P = .20 for Heterogeneity in relation to the intubation rate, no di-rectional trend in mortality was seen.
Risk Ratio (95% Confidence Interval) COMMENT
This systematic review and meta-
Nine of the studies found significant im- When the analysis was performed by analysis demonstrates the effectiveness provement in at least 1 of the main out- intention-to-treat, computing with- of noninvasive ventilation to reduce in- comes for which the trial was designed, drawals as events, RRs and 95% CIs tubation rate and mortality in patients whereas all the studies found sig- from a random-effects model did not with acute pulmonary edema. In a pre- nificant improvement in secondary differ significantly, showing a global vious systematic review published in reduction in mortality risk of 43% for 1998,38 CPAP was associated with a de- The analysis of the publication bias NIV (RR, 0.57; 95% CI, 0.41-0.79; crease in need for intubation (risk dif- yielded no significant results for either P = .73 for heterogeneity), which was ference −26%) and a trend to decrease test or comparison group.
46% for CPAP (RR, 0.54; 95% CI, mortality, but there was insufficient evi- 0.36-0.82; P = .40 for heterogeneity) dence on the effectiveness of NIPSV, and 37% for NIPSV (RR, 0.63; 95% CI, either compared with standard therapy NIV and Conventional Oxygen Ther-
0.37-1.06; P=.87 for heterogeneity).
or CPAP, because there were no ran- apy. Pooled data included 727 pa-
Taken together the 2 NIV modali- domized trials at that time. Neverthe- tients. Overall, NIV significantly re- ties demonstrated a significant 57% re- less, in the last 7 years, many studies have duced the risk of mortality compared duction in the need-to-intubate risk been published evaluating either CPAP with conventional oxygen therapy (P⬍.001; FIGURE 3). The decrease was
or NIPSV in patients with acute pulmo- (P⬍.001; FIGURE 2). The results were
statistically significant either for CPAP nary edema. Probably as a result of in- significant for CPAP, whereas NIPSV or NIPSV. Similar results were seen creased sample size, our meta-analysis in- tended toward a 40% reduction in the when the analysis was performed by in- cluding these trials has clearly reinforced risk of mortality (P = .07). However, tention to treat: 56% reduction in need- the role of CPAP in comparison with the number of patients studied with to-intubate risk for pooled data (RR, conventional therapy, showing a dra- NIPSV was lower than with CPAP and 0.44; 95% CI, 0.34-0.59; P=.31 for het- matic reduction in the need for intuba- the proportional weight for NIPSV in erogeneity), 60% reduction for CPAP tion (reduction in risk 60%) and a de- the pooled data analysis was only 35%.
(RR, 0.40; 95% CI, 0.28-0.58; P=.23 for crease in mortality (47%), which reached 3128 JAMA, December 28, 2005—Vol 294, No. 24 (Reprinted)
2005 American Medical Association. All rights reserved.
NONINVASIVE VENTILATION IN ACUTE CARDIOGENIC PULMONARY EDEMA statistical significance. Parallel to these Figure 4. Effects of Continuous Positive Airway Pressure vs Noninvasive Pressure Support
results, NIPSV demonstrated a similar re- duction in the need for intubation (52%)and a trend to decrease mortality in com- No. of Events/Total No.
Continuous Positive Noninvasive Pressure parison with conventional therapy. As in Support Ventilation the previous meta-analysis with CPAP,38 Source or Subcategory Mehta et al,35 1997 the impact of NIPSV on mortality did not Park et al,30 2001 reach statistical significance, possibly be- Bellone et al,36 2004 cause the number of patients included Crane et al,32 2004 in the model remains underpowered to Park et al,8 2004 Bellone et al,37 2005 demonstrate a substantial decrease inmortality. Although additional re- Risk Ratio, 0.9095% Confidence search would resolve this issue, current evidence on the effectiveness of NIV, es- P = .82P = .34 for Heterogeneity pecially CPAP, over conventional treat- Need to Intubate, No. of Events/Total No.
ment supports the use of this technique Source or Subcategory as standard therapy and further com- Mehta et al,35 1997 parisons between NIPSV and conven- Park et al,30 2001 tional oxygen therapy would not be con- Bellone et al,36 2004 sidered acceptable.
Crane et al,32 2004 In the comparison of NIV modali- Park et al,8 2004 Bellone et al,37 2005 ties, NIPSV has the potential advantageover CPAP of assisting the respiratory Risk Ratio, 1.4595% Confidence muscles during inspiration, which would result in faster alleviation of dyspnea and P = .39P = .63 for Heterogeneity exhaustion.12 Nevertheless, these physi- ological benefits did not translate into pri- Risk Ratio (95% Confidence Interval) mary outcomes in our meta-analysis,which did not find differences betweenCPAP and NIPSV in terms of intuba- technique offers advantage over the is less dependent on the experience or tion or mortality. This equivalence re- other and what subset of patients would the device and shows much lower vari- mained whether some nonpublished benefit more with either one of these ability in the studies. Third, besides ame- trials24,25 not included in the analysis, techniques remains unresolved.13 liorating fatigue, the main advantage of were incorporated into the model (data The present meta-analysis has sev- NIV is to avoid intubation and its asso- available on request). In addition, even eral limitations. First, criteria for diag- ciated complications, subsequently re- in patients with acute pulmonary edema nosis of acute pulmonary edema are not ducing the mortality rate. Several stud- and hypercapnia, a condition usually as- well established. In the new guidelines ies used rescue NIV, sometimes NIPSV sociated with muscle fatigue, a recent on the diagnosis and treatment of acute in CPAP groups or either NIPSV or study did not demonstrate differences be- heart failure proposed by the Euro- CPAP in conventional groups. This tween these techniques either.37 Hyper- pean Society of Cardiology,40 2 types of might have introduced some conserva- capnic patients were expected to be the acute pulmonary edema are recog- tive bias in the estimation of the mor- target population for NIPSV for physi- nized: hypertensive crisis and non– tality rate. Fourth, although our analy- ological reasons and especially after the hypertensive pulmonary edema. The sis did not find significant publication favorable results in the post hoc analy- prognosis in terms of intubation and bias, this result must be taken with cau- ses of some studies using NIPSV.34,39 mortality differs41 and the proportion of tion due to the low power of tests ana- The incidence of myocardial infarc- each type of acute pulmonary edema in- lyzing this issue when the number of tion for the interventional therapies ana- cluded in the studies was not well de- trials is small. Finally, although many lyzed in the studies was similar. Al- fined. Second, the characteristics of the trials of this meta-analysis included a though a preliminary study35 described ventilators (displays, leakage compen- small sample size, more than half were a higher rate of acute myocardial infarc- sation, FiO2 range, trigger, etc), the level powered enough to demonstrate signifi- tion with NIPSV, no other trial found of NIPSV used and the experience of the cant differences between interventions this incidence and a recent study, spe- teams were relatively different in the in the main outcomes. The limited size cifically addressing this issue, showed no trials and all of these variables may in- of some of these trials, however, rein- differences between both techniques.36 fluence the results of this technique.42 forces the necessity of our meta- Therefore, the question of whether one This is not the case for CPAP because it analysis. In addition, it should be men- 2005 American Medical Association. All rights reserved.
(Reprinted) JAMA, December 28, 2005—Vol 294, No. 24 3129
NONINVASIVE VENTILATION IN ACUTE CARDIOGENIC PULMONARY EDEMA tioned that the critical phase of acute 2. Ram FS, Picot J, Lightowler J, Wedzicha JA. Non-
acute cardiogenic pulmonary edema. J Formos Med invasive positive pressure ventilation for treatment of pulmonary edema, when patients are eli- respiratory failure due to exacerbations of chronic ob- 24. Bollaert PE, Sauder PH, Girard F, et al. Continuous
gible, may be extremely short because structive pulmonary disease. Cochrane Database Syst positive airway pressure (CPAP) vs proportional assist some patients may require immediate in- ventilation (PAV) for noninvasive ventilation in cardio- 3. Ra¨sa¨nen J, Heikkla¨ J, Downs J, Nikki P, Va¨isa¨nen I,
genic pulmonary edema (CPE). Am J Respir Crit Care tubation or may rapidly ameliorate af- Viitanen A. Continuous positive airway pressure by face Med. 2002;165(suppl 8):A387.
ter starting medical therapy. This rapid mask in acute cardiogenic pulmonary edema. Am J 25. Liesching TN, Cromier K, Nelson D, Short K, Su-
cov A, Hill NS. Bilevel noninvasive ventilation vs con- evolution has seriously limited the re- 4. Bersten AD, Holt AW, Vedig AE, Skowronski GA, Bag-
tinuous positive airway pressure to treat acute pulmo- cruitment capacity of the studies.
goley CJ. Treatment of severe cardiogenic pulmonary nary edema. Am J Respir Crit Care Med. 2003;167(suppl edema with continuous positive airway pressure deliv- Despite these limitations our quan- ered by face mask. N Engl J Med. 1991;325:1825-1830.
26. Mackay CA, Mackay TW, Barr K, et al. Random-
titative systematic review of existing lit- 5. Lin M, Yang YF, Chiang HT, Chang MS, Chiang BN,
ized controlled trial of CPAP vs conventional therapy in Cheitlin MD. Reappraisal of continuous positive air- acute pulmonary edema. Am J Respir Crit Care Med.
erature demonstrates that NIV re- way pressure therapy in acute cardiogenic pulmonary duces intubation rate and mortality in edema. Chest. 1995;107:1379-1386.
27. Hao CX, Luo XR, Liu YM. Treatment of severe car-
6. Takeda S, Nejima J, Takano T, et al. Effect of nasal
diogenic pulmonary edema with continuous positive air- patients with acute pulmonary edema.
continuous positive airway pressure on pulmonary edema way pressure by basal face mask. Acta Academiae Me- Noninvasive ventilation has recently complicating acute myocardial infarction. Jpn Circ J.
dicinae Jiangxi. 2002;42:48-50.
28. Moritz F, Benichou J, Vanheste M, et al. Boussig-
been categorized as class IIa, level of evi- 1998;62:553-558.
7. L'Her E, Duquesne F, Girou E, et al. Noninvasive con-
nac continuous positive airway pressure device in the dence A, in the guidelines on the diag- tinuous positive airway pressure in elderly cardiogenic emergency care of acute cardiogenic pulmonary oe- nosis and treatment for acute heart pulmonary edema patients. Intensive Care Med. 2004; dema: a randomized pilot study. Eur J Emerg Med. 2003; failure by the European Society of Car- 8. Park M, Sangean MC, Volpe MC, et al. Random-
29. Takeda S, Takano T, Oqawa R. The effect of nasal
diology,40 based on some of the trials ized, prospective trial of oxygen, continuous positive air- continuous positive airway pressure on plasma endo- way pressure, and bilevel positive airway pressure by thelin-1 concentrations in patients with severe cardio- analyzed in the present study.3-6,9,22,31,35 face mask in acute cardiogenic pulmonary edema. Crit genic pulmonary edema. Anesth Analg. 1997;84:1091- Given the results of our review, we Care Med. 2004;32:2407-2415.
9. Masip J, Betbese´ AJ, Pa´ez J, et al. Non-invasive pres-
30. Park M, Lorenzi-Filho G, Feltrim MI, et al. Oxygen
think that NIV should be strongly con- sure support ventilation versus conventional oxygen therapy, continuous positive airway pressure, or non- sidered as a first-line treatment.
therapy in acute cardiogenic pulmonary edema: a ran- invasive bilevel positive pressure ventilation in the treat- domized study. Lancet. 2000;356:2126-2132.
ment of acute cardiogenic pulmonary edema. Arq Bras Author Contributions: Dr Masip had full access to all
10. Antonelli M, Conti G, Moro ML, et al. Predictors
of the data in the study and takes responsibility for of failure of noninvasive positive ventilation in patients 31. Kelly CA, Newby DE, McDonagh TA, et al. Ran-
the integrity of the data and the accuracy of the data with acute hypoxemic respiratory failure. Intensive Care domised controlled trial of continuous positive airway pressure and standard oxygen therapy in acute pulmo- Study concept and design: Masip, Roque, Sa´nchez, 11. Burns KE, Sinuff T, Adhhikari NK, et al. Bilevel non-
nary oedema. Eur Heart J. 2002;23:1379-1386.
invasive positive pressure ventilation for acute respira- 32. Crane SD, Elliott MW, Gilligan P, Richards K, Gray
Acquisition of data: Masip, Roque, Sa´nchez, Subirana, tory failure. Crit Care Med. 2005;33:1477-1483.
AJ. Randomised controlled comparison of continuous 12. Wysocki M. Noninvasive ventilation in acute car-
positive airways pressure, bilevel non-invasive ventila- Analysis and interpretation of data: Masip, Roque, diogenic pulmonary edema: better than continuous posi- tion, and standard treatment in emergency depart- Sa´nchez, Subirana, Expo´sito tive airway pressure? Intensive Care Med. 1999;25:1-2.
ment in patients with acute cardiogenic pulmonary Drafting of the manuscript: Masip.
13. Mehta S, Nava S. Mask ventilation and cardio-
oedema. Emerg Med J. 2004;21:155-161.
Critical revision of the manuscript for important in- genic pulmonary edema. Intensive Care Med. 2005; 33. Levitt MA. A prospective, randomized trial of BIPAP
tellectual content: Masip, Roque, Sa´nchez, Ferna´ndez, in severe acute congestive heart failure. J Emerg Med.
14. Jadad AR, Moore RA, Carroll D, et al. Assessing the
quality of reports of randomized clinical trials: is blind- 34. Nava S, Carbone G, Dibatista N, et al. Noninva-
Statistical analysis: Masip, Roque.
ing necessary? Control Clin Trials. 1996;17:1-12.
sive ventilation in cardiogenic pulmonary edema. Am J Obtained funding: Masip, Roque.
15. Schulz KF, Chalmers I, Hayes RJ, Altman D. Em-
Respir Crit Care Med. 2003;168:1432-1437.
Administrative, technical, or material support: Roque, pirical evidence of bias. JAMA. 1995;273:408-412.
35. Mehta S, Jay GD, Woolard RH, et al. Random-
Sa´nchez, Ferna´ndez, Subirana, Expo´sito.
16. Egger M, Smith GD, Schneider M, Minder C. Bias
ized, prospective trial of bilevel vs continuous positive Study supervision: Masip, Roque, Ferna´ndez.
in meta-analysis detected by a simple, graphical test.
airway pressure in acute pulmonary edema. Crit Care Financial Disclosures: None reported.
Funding/Support: This project was funded in part by
17. Begg CB, Mazumdar M. Operating characteristics
36. Bellone A, Monari A, Cortellaro F, Vettorello M, Ar-
the nonprofit Agencia de Evaluacio´n de Tecnologı´as of a rank correlation test for publication bias. Biometrics.
lati S, Coen D. Myocardial infarction rate in acute pul- Sanitarias, Instituto de Salud Carlos III, Fondo de In- monary edema. Crit Care Med. 2004;32:1860-1865.
vestigacio´n Sanitaria (FIS) PI04/90064. The grant cov- 18. Vaisanen IT, Rasanen J. Continuous positive air-
37. Bellone A, Vettorello M, Monari A, Cortellaro F, Coen
ered administrative material and a scholarship.
way pressure and supplemental oxygen in the treat- D. Noninvasive pressure support ventilation vs continu- Role of the Sponsor: Agencia de Evaluacio´n de Tec-
ment of cardiogenic pulmonary edema. Chest. 1987;92: ous positive airway pressure in acute hypercapnic pul- nologı´as Sanitarias, Instituto de Salud Carlos III, Fondo monary edema. Intensive Care Med. 2005;31:807-811.
de Investigacio´n Sanitaria, did not participate in the 19. Chadda K, Annane D, Hart N, et al. Cardiac
38. Pang D, Keenan SP, Cook DJ, Sibbald WJ. The effect
design and conduct of the study, in the collection, and respiratory effects of continuous positive airway of positive pressure airway support on mortality and need analysis, and interpretation of the data, or in the prepa- pressure and noninvasive ventilation in acute cardiac for intubation in cardiogenic pulmonary edema: a sys- ration, review, or approval of the manuscript.
pulmonary edema. Crit Care Med. 2002;30:2457- tematic review. Chest. 1998;114:1185-1192.
Acknowledgment: We thank Carolyn Newey (lan-
39. Rusterholtz T, Kempf J, Berton C, et al. Noninva-
guage editor at Institut de Recerca de l'Hospital de la 20. Craven RA, Singletary N, Bosken L, et al. Use of bi-
sive pressure support ventilation (NIPSV) with face mask Santa Creu I Sant Pau, Barcelona) and Mitsi Ito (En- level positive airway pressure in out-of-hospital patients.
in patients with acute pulmonary edema (ACPE). In- glish Essentials) for editing the manuscript and Joa- Acad Emerg Med. 2000;7:1065-1068.
tensive Care Med. 1999;25:21-28.
quim Pa´ez (attending physician, ICU Department, Hos- 21. Cross AM, Cameron P, Kierce M, Ragg M, Kelly AM.
40. Nieminen MS, Bohm M, Cowie MR, et al. Execu-
pital Dos de Maig de Barcelona) for his constant support Non-invasive ventilation in acute respiratory failure.
tive summary of the guidelines on the diagnosis and throughout the project.
Emerg Med J. 2003;20:531-534.
treatment of acute heart failure. Eur Heart J. 2005;26: 22. Sharon A, Shpirer I, Kaluski E, et al. High-dose in-
travenous isosorbide-dinitrate is safer and better than 41. Masip J, Pen˜a C, Figueras J, Pa´ez J, Sa´nchez B, Can-
Bi-PAP ventilation combined with conventional treat- cio B. Hypertensive acute pulmonary edema. Eur J Heart 1. Brochard L, Mancebo J, Wysocki M, et al. Nonin-
ment for severe pulmonary edema. J Am Coll Cardiol.
Fail. 2005;4(suppl 1):11.
vasive ventilation for acute exacerbations of chronic ob- 42. Masip J, Pa´ez J, Betbese´ AJ, Vecilla F. Noninvasive
structive pulmonary disease. N Engl J Med. 1995;333: 23. Lin M, Chiang HT. The efficacy of early continu-
ventilation for pulmonary edema in emergency room.
ous positive airway pressure therapy in patients with Am J Respir Crit Care Med. 2004;169:1072-1073.
3130 JAMA, December 28, 2005—Vol 294, No. 24 (Reprinted)
2005 American Medical Association. All rights reserved.

Source: http://www.saladeaula.abitep.com.br/upload/slidesoff/vd_744f/revis%C3%A3o%20sistematica%20e%20meta%20an%C3%A1lise.pdf