Evidence Report/Technology Assessment

Number 23

Prepared for:
Agency for HealthCare Research and Quality

U.S. Department of Health and Human Services
2101 East Jefferson Street
Rockville, MD 20852
http://www.ahrq.gov

Contract No. 290-97-0017

Prepared by:
Department of Medicine and Clinical Epidemiology & Biostatistics
McMaster University, Hamilton, Ontario, Canada
Deborah Cook, MD
Principal Investigator
Maureen Meade, MD
Gordon Guyatt, MD
Lauren Griffith, MSc
Lynda Booker, BA
Investigators

AHRQ Publication No. 01-E010

November 2000
ISBN 1-58763-011-7
ISSN 1530-4396

On December 6, 1999, under Public Law 106-129, the Agency for Health Care Policy and Research (AHCPR) was reauthorized and renamed the Agency for Healthcare Research and Quality (AHRQ). The law authorizes AHRQ to continue its research on the cost, quality, and outcomes of health care, and expands its role to improve patient safety and address medical errors.

This report may be used, in whole or in part, as the basis for development of clinical practice guidelines and other quality enhancement tools, or a basis for reimbursement and coverage policies. AHRQ or U.S. Department of Health and Human Services endorsement of such derivative products may not be stated or implied.

The Agency for Healthcare Research and Quality (AHRQ), through its Evidence-based Practice Centers (EPCs), sponsors the development of evidence reports and technology assessments to assist public- and private-sector organizations in their efforts to improve the quality of health care in the United States. The reports and assessments provide organizations with comprehensive, science-based information on common, costly medical conditions and new health care technologies. The EPCs systematically review the relevant scientific literature on topics assigned to them by AHRQ and conduct additional analyses when appropriate prior to developing their reports and assessments.

To bring the broadest range of experts into the development of evidence reports and health technology assessments, AHRQ encourages the EPCs to form partnerships and enter into collaborations with other medical and research organizations. The EPCs work with these partner organizations to ensure that the evidence reports and technology assessments they produce will become building blocks for health care quality improvement projects throughout the Nation. The reports undergo peer review prior to their release.

AHRQ expects that the EPC evidence reports and technology assessments will inform individual health plans, providers, and purchasers as well as the health care system as a whole by providing important information to help improve health care quality.

We welcome written comments on this evidence report. They may be sent to: Director, Center for Practice and Technology Assessment, Agency for Healthcare Research and Quality, 6010 Executive Blvd., Suite 300, Rockville, MD 20852.

John M. Eisenberg, M.D.	Douglas B. Kamerow, M.D.
Administrator	Director, Center for Practice and Technology Assessment
Agency for Healthcare Research and Quality	Agency for Healthcare Research and Quality

The authors of this report are responsible for its content. Statements in the report should not be construed as endorsement by the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services of a particular drug, test, treatment, or other clinical service.

We would like to thank the following individuals who helped with the data abstraction for this systematic review: Aziz Aldawood, Ron Butler, Brad Elms, Lori Hand, Andie Ingram, Carmen Kergl, Jill Randall, Tasnim Sinuff, and Gemini Toprani. Our appreciation is extended to Heiner Bucher, Toshio Fukuoka, Luz Letelier, and Marco Ranieri for the translation of, or data abstraction from, foreign language articles. We also thank Ann McKibbon for her expertise with database searching; Bruce Weaver, Dr. Stephen Walter, and Robin Roberts for their help with the analysis; Marlene Taylor for her help with the preparation of this report; Anne Snider for her administrative guidance; and Drs. Sydney Parker and Alejandro Jadad for their support over the course of this review. We are grateful for the expert peer review suggestions from Drs. Laurent Brochard, Suzanne Burns, Wesley Ely, Scott Epstein, Jesse Hall, John Heffner, Dean Hess, Rolf Hubmayr, Bob Kacmarek, Neil McIntyre, and Peter Pronovost; their suggestions have helped to shape this document. Finally, we appreciate the input of our scientific advisors Drs. Anne Perry and Arthur Slutsky for their expertise and suggestions.

Because mechanical ventilation incurs significant morbidity, mortality, and costs and because premature extubation as well as delayed extubation can cause harm, weaning that is both expeditious and safe is highly desirable. We sought to determine whether the current literature elucidates when and how weaning should begin, proceed, and end.

We used five computerized bibliographic databases, hand searching, bibliographic references, expert consultation, and a duplicate independent review process to identify relevant articles.

Our eligibility criteria were broad. We selected studies evaluating any weaning modes, algorithms, or other interventions to facilitate weaning; studies of weaning predictors; and studies evaluating patient and nursing experiences during weaning.

We developed generic forms to abstract data from all studies, in addition to forms specific to randomized trials, nonrandomized controlled studies, and studies of weaning predictors. We developed an instruction manual and trained eight individuals to abstract data related to study characteristics and results using duplicate, independent review. Quantitative data were abstracted using several metrics. We pooled results across randomized trials and across studies of weaning predictors only when our assessment of the patients, interventions, and outcomes indicated that pooling was legitimate.

We reviewed 154 articles. For stepwise reductions in mechanical support, pressure support mode or multiple daily T-piece trials may be superior to synchronized intermittent mandatory ventilation. For trials of unassisted breathing, low levels of pressure support may be beneficial. There may be substantial benefits to early extubation and institution of noninvasive positive pressure ventilation before patients are ready to breathe without mechanical assistance. The value of differing modes as reflected in these studies depends on the thresholds for initiating, progressing through, and terminating weans in the specific study protocols. Unfortunately, these thresholds involve more than objective data and appear to be related to physician judgment.

The implementation of nurse-driven or respiratory therapist-driven weaning protocols, regardless of what modes are employed, significantly expedites weaning and is probably safe. Following cardiac surgery, early extubation is unequivocably achieved with a variety of anesthetic interventions and intensive care unit protocols; however, the corresponding reduction in intensive care unit stay is generally small and the impact on complications, though rare, remains unclear. The role of computerized protocols has not been established.

We did not uncover any consistently powerful weaning predictors. The most frequently studied and one of the most helpful tests is the rapid shallow breathing index; however, the pooled likelihood ratio for a positive test ranged from 1.3 to 2.8. Two other predictors, occlusion pressure/maximum inspiratory pressure and the compliance, rate, oxygenation, and pressure index, are more powerful, though less intensively studied. In general, the probable reason for the poor performance of weaning predictors is that physicians have already considered the results of these predictors when they select patients for study.

Future research initiatives should include: determining the optimal tradeoff between prolonged time on a ventilator and reintubation in specific patient groups, further evaluation of weaning protocols (What types of patients are most likely to benefit? Which protocols are most effective? How large are the associated cost reductions? Is there a role for computers?), and clarification of the risk-benefit of early extubation with noninvasive positive pressure ventilation.

This document is in the public domain and may be used and reprinted without permission except those copyrighted materials noted for which further reproduction is prohibited without the specific permission of copyright holders.

Cook D, Meade M, Guyatt G, et al. Criteria for Weaning From Mechanical Ventilation. Evidence Report/Technology Assessment No. 23 (Prepared by McMaster University under Contract No. 290-97-0017). AHRQ Publication No. 01-E010. Rockville MD: Agency for Healthcare Research and Quality. November 2000.

The majority of critically ill patients in most modern intensive care units (ICUs) require a period of mechanical ventilation. Prolonged mechanical ventilation is associated with nosocomial pneumonia, cardiac-associated morbidity, and death. Prematurely discontinuing mechanical ventilation may result in reintubation which is associated with complications similar to prolonged ventilation. Thus, optimal weaning-minimizing the duration of mechanical ventilation without incurring substantial risk of reintubation and thus preventing important complications-is a crucial part of the management of critically ill patients. Therefore, the Agency for Healthcare Research and Quality (AHRQ) asked us to address the following questions:

We resolved to retrieve all randomized trials and the most relevant clinical observational studies that addressed the AHRQ review questions. Conceptually, we were interested in any patients receiving mechanical ventilation, in any strategies designed to facilitate weaning and extubation, in predictors of weaning and extubation in all critically ill patients, and in predictors of the duration of weaning in patients with chronic obstructive pulmonary disease (COPD) or patients following cardiac surgery.

We included all studies of adult and pediatric patients who were mechanically ventilated and had either an endotracheal tube or tracheostomy tube. We excluded studies of highly specific populations and studies in neonates.

We included studies conducted in intensive care units (ICUs), intermediate care units, stepdown units, and postanesthetic recovery rooms. We excluded studies of home ventilation for children or adults and chronic ventilation settings.

We included any ventilation or weaning strategy (e.g., mode, method, procedure, protocol, timing, operator, computer, tracheostomy, noninvasive ventilation modes, adjunctive holistic aids, and other miscellaneous approaches) geared to facilitate weaning and/or extubation. We excluded interventions focused on mechanical ventilation methods and interventions whose influence on the duration of ventilation has already been summarized in a recent systematic review (e.g., sedation in the ICU and optimal timing of tracheotomy).

We included predictors of weaning and/or extubation success and predictors of duration of mechanical ventilation in cardiac surgery and COPD patients. We excluded predictors of self-extubation.

We included all clinical outcomes. We excluded studies which reported exclusively physiologic outcomes. We included studies reporting the endpoint of patients' experiences and nurses' assessments of patients' experiences to address question 4 regarding the important role of nurses in optimizing the weaning process.

To identify relevant studies, we searched MEDLINE, EMBASE, HealthSTAR, CINAHL, the Cochrane Controlled Trials Registry, and the Cochrane Data Base of Systematic Reviews from 1971 to 1998. We also examined reference lists and personal files and hand searched Respiratory Care. We did not explicitly search for unpublished literature. We retrieved all articles that either of two reviewers of the titles and abstracts considered possibly eligible. The same two reviewers examined the full text and made final decisions regarding eligibility based on the inclusion and exclusion criteria described above.

Five respiratory therapists and five intensivists participated in data abstraction and in rating the methodologic quality of all eligible randomized trials or nonrandomized controlled cohort studies addressing treatment issues, and all studies providing quantitative data concerning predictors of weaning and extubation. Two reviewers abstracted the data and assessed the methodologic quality of each study.

Methodologic features of randomized trials that we abstracted included the method of randomization and whether randomization was concealed; the extent to which groups were similar with respect to important prognostic factors; whether investigators conducted an intention to treat analysis; whether patients, clinicians, and those assessing outcome were blind to allocation; the extent to which the groups received similar cointerventions; and the reporting of the reasons for study withdrawal.

For nonrandomized controlled clinical trials, we considered the extent to which groups were similar with respect to important prognostic factors, whether the investigators adjusted for differences in prognostic factors, and the extent to which the groups received similar cointerventions.

For studies addressing predictors of weaning success, we considered whether investigators enrolled a representative sample of patients and whether those making weaning decisions or assessing outcomes were blind to predictor variables.

For qualitative studies, we considered whether the choice of participants was relevant to the research question and if their selection was well reasoned, whether the data collection methods were appropriate for the research objectives, whether the data collection was comprehensive enough to support rich and robust descriptions of the observed events, and whether the data were appropriately analyzed and the findings adequately corroborated.

To synthesize the data from randomized trials comparing weaning interventions we abstracted or, when necessary, calculated effect sizes in terms of relative risks and associated 95 percent confidence intervals (CIs) for binary outcomes and mean differences and 95 percent CIs for continuous variables. We reviewed the interventions and outcomes and decided when it was legitimate to pool across studies and when it was not. When pooling was not appropriate, we categorized studies according to similarity of interventions. All our pooled analyses are based on a random effects model that includes differences between studies in calculating the variance estimate provided the strategy for final estimates of all treatment effects. For nonrandomized studies that compared alternative weaning interventions, we used similar methodology for calculating point estimates and CIs for individual studies but made no attempt to pool data across studies.

For observational studies addressing prediction of successful weaning and duration of ventilation, we categorized studies according to the outcome of interest and, for each predictor in each study in which the data were available, we constructed a 2X2 table examining the presence or absence of the predictor in relation to the success or failure of the weaning process. This procedure allowed calculation of the sensitivity and specificity of the tests and their associated 95 percent CIs as well as odds ratios and their 95 percent CIs and the associated likelihood ratios. We then organized the observational studies according to predictors of interest. We defined predictors as relevant if they showed potential for differentiating success from failure and retained all predictors for which results were presented in 2X2 tables if there was an associated likelihood ratio (LR) of greater than 2 or lower than 0.5. When results were presented as means and standard deviations of the success and failure groups, we included predictors if the difference in means between the two groups was greater than one-half of the smaller of the standard deviations of the two groups. Where appropriate, we pooled the observational data to narrow the 95 percent CIs.

• We reviewed 154 articles.
• The issue of the optimal start of weaning is confounded by alternative definitions of weaning: one reasonable conceptualization is weaning beginning with the onset of mechanical ventilation. Research to date suggests the best answer to "when to start weaning" is to develop a protocol implemented by nurses and respiratory therapists that begins testing for the opportunity to reduce support very soon after intubation and reduces support at every opportunity.
• Differences in clinicians' intuitive threshold for reduction or discontinuation of ventilatory support have a greater impact on failure of spontaneous breathing trials, or on reintubation, than do modes of weaning. When clinicians set a high threshold, many patients who could tolerate weaning remain on mechanical support longer than is necessary.
• For stepwise reductions in mechanical support, pressure support mode or multiple daily T-piece trials may be superior to intermittent mandatory ventilation.
• For trials of unassisted breathing, low levels of pressure support may be beneficial.
• There may be substantial benefits to early extubation and institution of noninvasive positive pressure ventilation for patients who are alert, cooperative, and are ready to breathe without an artificial airway.
• Following cardiac surgery, early extubation is unequivocably achieved with a variety of anesthetic interventions and ICU protocols; however, the corresponding reduction in ICU stay is generally small and the impact on complications, though rare, remains unclear.
• The role of computerized weaning protocols has not been established.
• Although steroids can reduce postextubation stridor in children, their impact on reintubation in children and adults remains uncertain.
• Most theoretically plausible predictors of weaning and extubation success have no predictive power. Those with some predictive power include the rapid shallow breathing index, which has been most intensively studied, and occlusion pressure/maximum inspiratory pressure (P_0.1/MIP) and the compliance, rate, oxygenation, and pressure (CROP) index. However, these are relatively weak predictors of weaning success. Tests are rarely useful in increasing the probability of weaning success; on occasion, they can lead to moderate reductions in the probability of success. In general, weaning predictors were probably found to perform poorly because physicians have already considered the results when they select patients for study.

• Examination of alternative weaning strategies should enroll homogeneous patient groups: those whose likely period of additional ventilation is a few hours, and those whose likely period is a few days. Patients after cardiac surgery constitute another population that should be considered separately.
• In the setting of a high threshold for extubation associated with low failure rates, investigators would require trials of thousands of patients to demonstrate differences between techniques and tens of thousands to demonstrate differences in complications of failed extubation. Investigators should establish plausible event rates before embarking on clinical trials.
• Investigators should attempt to elucidate the tradeoff between decreasing duration of time on a ventilator and the increase in reintubation rates associated with a low weaning threshold (e.g., what reduction in duration of time on a ventilator would warrant an increase in reintubation rates from 5 to 10 percent)? This work should involve attention to the important consequences of prolonged ventilation or reintubation, including nosocomial pneumonia, cardiac morbidity, and death.
• Investigators should launch trials examining the use of noninvasive positive pressure ventilation (NPPV) in reducing the duration of intubation and total mechanical support. Future research should also explore the optimal timing and management of NPPV for weaning purposes, its effect on morbidity (e.g., pneumonia), length of ICU stay, and mortality.
• Investigators should launch additional randomized trials of weaning protocols implemented by respiratory therapists and nurses. These trials should evaluate the differential impact of protocols in different types of patients and in ICUs with different organizational structures (e.g., open versus closed units, teaching versus community hospitals). The influence of different protocols and their impact on ICU and hospital length of stay and costs are important future considerations.
• A more fruitful line of investigation than further research seeking powerful predictors of successful weaning or extubation might be randomized trials of weaning protocols that decrease the duration of mechanical ventilation without substantially increasing rates of failed extubation.