© 2004 American Thoracic Society
Noninvasive Ventilation for Persistent Weaning FailureTo the Editor:The study by Ferrer and colleagues (1) could have a significant impact on the care, cost of care, and outcome of patients weaning from mechanical ventilation. The study, however, raises important concerns about its conduct, validity, and interpretation. First, the patients weaned by noninvasive ventilation (NIV) were extubated despite having a frequency-to–tidal volume ratio greater than 105, putting them at a very high risk for failing the extubation attempt (2). How did the investigators explain this increased risk to the patients or their surrogate decision makers during the process of obtaining informed consent? Second, the investigators excluded 23 eligible patients before randomization, introducing a potential for selection bias. Moreover, they did not describe how NIV was "gradually withdrawn," who was responsible for its administration (nurses, respiratory therapists, or physicians), or what was their level of training in NIV. Furthermore, they did not describe the level of staffing at the units where NIV was administered or whether patients participating in the study were cared for differently from patients that did not participate. The labor-intensive nature of NIV could have inadvertently led to a higher level of care, resulting in a better outcome. Data about the 23 excluded patients and more details about the weaning team and weaning protocol in the NIV arm are important to give the study findings external validity. Finally, why was weaning with NIV better than weaning with a daily trial of spontaneous breathing? Did the standard method of weaning based on extubating patients that have an adequate breathing pattern and can tolerate a 2-hour spontaneous breathing trial fail to identify patients ready for extubation? Or was it the reduced incidence of nosocomial pneumonias in the NIV arm? In the conventional weaning arm, nosocomial pneumonias could have contributed to, rather than just complicating, the longer weaning process if they occurred earlier rather than later after randomization. The investigators should provide data on when the pneumonias occurred in the conventional weaning arm to show whether the pneumonias were a cause or just a result of the prolonged mechanical ventilation.
Cook County Hospital/Rush University Medical Center Chicago, Illinois FOOTNOTES Conflict of Interest Statement: A.T. and B.M. have no declared conflict of interest. REFERENCES
From the Authors: Drs. Tulaimat and Mokhlesi have raised several concerns regarding our recent publication (1). The frequency-to–tidal volume (f/VT) ratio is probably the best individual predictor of weaning failure (2). However, whether the presence of an f/VT above a predefined threshold is itself a contraindication for extubation is not supported by any evidence; indeed, a recent systematic revision from experts in this field found a modest accuracy of the individual indexes in predicting weaning outcome (3). The rationale for applying noninvasive ventilation (NIV) in these patients was clearly stated in the article. Patients failing weaning attempts often develop a rapid and shallow breathing pattern (4), and NIV allows the correction of such an abnormal breathing pattern (5). Regarding excluded patients, we agree that this may introduce a potential for selection bias. The presence of exclusion criteria is a limitation inherent in all such controlled clinical trials, especially in severely ill awake patients in whom several features need to be under control. However, the concealment of randomization before patients are enrolled minimizes such bias. Most of the requested information regarding the weaning protocol is available in the online data supplement; we recommend to Drs. Tulaimat and Mokhlesi a more careful reading of this section. We progressively reduced both the levels of inspiratory support and the daily duration of NIV. Regarding the attending team, the standard staffing of Spanish intensive care units (ICUs) is one nurse for every two patients, whereas our unit has full-time respiratory therapists 14 hours daily, senior respiratory and critical care physicians together with training physicians, and physicians on call 24 hours daily. NIV has been routinely used in our units since the beginning of the 1990s; hence, the learning curve for NIV had been achieved when the protocol was initiated. The level and quality of care in enrolled patients was, except for the specific interventions of the protocol, the same as for the remaining patients admitted in the ICUs. Indeed, the consumption of human resources of NIV is not substantially different than invasive ventilation (6). The reasons for the better outcome of the NIV group are also widely discussed in the article, namely the earlier extubation of patients taking advantage of NIV compared with the conventional arm. Only episodes of nosocomial pneumonia occurring after enrollment of patients were reported; this is clearly stated in the article. The relationship between nosocomial pneumonia and prolonged ventilation is reciprocal; prolonged invasive ventilation is a risk factor for nosocomial pneumonia, and nosocomial pneumonia is a cause of prolonged ventilation.
Institut Clínic de Pneumologia i Cirurgia Toracica Barcelona, Spain FOOTNOTES Conflict of Interest Statement: M.F. and A.T. have no declared conflict of interest. REFERENCES
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
