High Mortality in Patients with COPD Exacerbations Who Fail Noninvasive Ventilation

Abstract & Commentary

By David J. Pierson, MD, Professor Emeritus, Pulmonary and Critical Care Medicine, University of Washington, Seattle

This article originally appeared in the February 2012 issue of Critical Care Alert. It was peer reviewed by William Thompson, MD. Dr. Thompson is Associate Professor of Medicine, University of Washington, Seattle. Drs. Pierson and Thompson report no financial relationships relevant to this field of study.

Synopsis: Wider use of noninvasive ventilation in managing severe COPD exacerbations has improved overall outcomes, but this study of a large nationwide database shows that increasing numbers of patients fail NIV and require intubation. This subset of patients has substantially higher mortality and hospital costs.

Source: Chandra D, et al. Outcomes of non-invasive ventilation for acute exacerbations of COPD in the United States, 1998-2008. Am J Respir Crit Care Med 2011;Oct 20. [Epub ahead of print.]

Until now, getting a handle on the utilization and outcomes of noninvasive ventilation (NIV) in the management of acute exacerbations of chronic obstructive pulmonary disease (AECOPD) in the United States has been hampered by the absence of a nationwide database on this important condition.1 With this study, Chandra and colleagues have gone a long way toward correcting this deficiency. They used a nationwide database to examine more than 7 million hospitalizations for AECOPD between 1998 and 2008, looking at patient demographics, the use of both NIV and invasive mechanical ventilation (IMV), and patient outcomes. Although the article provides a general overview of their findings, I will focus primarily on what their data reveal about an important subset of AECOPD patients — those in whom NIV is unsuccessful, necessitating intubation and the use of IMV.

The authors used data from the Nationwide Inpatient Sample of the Healthcare Cost and Utilization Project (HCUP-NIS), which collects information from about 20% of all U.S. hospitals and includes all areas of the country, insurance categories, and hospital sizes. For all admissions for AECOPD, Chandra et al determined whether NIV or IMV had been used and identified all patients who were started on NIV and subsequently received IMV; against these data they examined in-hospital mortality (adjusted for patient demographics, hospital characteristics, and comorbidities), length of stay, and total charges for the hospitalization.

Of the 7,511,267 admissions for AECOPD to the participating hospitals from 1998 through 2008, 612,650 (8.1%) received ventilatory support, with NIV progressively increasing (from 1.0% to 4.5% of admissions) and IMV decreasing (from 6.0% to 3.5%) during the study period. The proportion of patients who started on NIV and were switched to IMV remained the same at about 4.6%, but because of the steady increase in patients in whom NIV was used, their absolute numbers climbed steadily. Overall, 9681 patients transitioned from NIV to IMV, of whom 2595 (27%) died. This compares to 17,436 deaths (9%) among 198,375 patients managed initially on NIV who were not switched to IMV. Adjusted odds ratios for mortality in NIV patients who were switched to IMV were substantially increased in comparison with those who remained on NIV, for every year starting in 2000. In 2008, a patient requiring IMV after unsuccessful NIV had 61% greater odds of death than a patient placed directly on IMV at the start, and 677% greater odds of death compared to a patient treated with NIV without transition to IMV. Charges for hospitalization increased steadily from 1998 through 2008, but the increase was steepest for patients who required a transition from NIV to IMV. Hospital length of stay was longest for the latter group and did not fall over time, whereas length of stay gradually decreased among the other groups.


This study demonstrates that whether ventilatory support was required was strongly correlated with outcomes among patients hospitalized with AECOPD. The mortality rate among those in whom neither NIV nor IMV was used remained in the 2-3% range throughout the 11-year data period. For patients who received NIV only, in-hospital mortality started at 11 or 12% and fell gradually to 7% or 8% (estimates from Figure 3, since precise data are not provided in the paper). Patients who received only IMV (no NIV) had an overall mortality rate of 23%, whereas, as mentioned, 27% of the patients who started out on NIV and had to be intubated for IMV died. Of course, these groups of patients undoubtedly varied a great deal in terms of severity of underlying COPD, the acute illness, comorbidities, and other factors, so that the mortality differences cannot be interpreted as simply reflecting differences in ventilatory support. Still, I think some conclusions can be drawn.

The steady increase in NIV use nationwide during the study period probably reflects increased awareness of the compelling evidence supporting it and the recommendations of practice guidelines, and the progressive decrease in mortality among patients who did not require intubation is consistent with increasing experience and expertise with this therapy over time. That the proportion of patients receiving IMV fell by nearly half during the study period likely also reflects increasing awareness that this complication-laden intervention is not necessary as often as we used to believe; the mortality rate among this decreasing proportion of patients who required IMV stayed about the same.

The group of greatest concern, whose outcomes actually worsened over time in terms of absolute patient numbers, is those patients who were initially managed with NIV and subsequently required intubation. Some of these patients likely developed new complications or experienced a progression of the primary problem. However, it is also likely that others were unsuitable for NIV in the first place, or received NIV in an ineffective or suboptimal manner. Carrying out NIV effectively is both a science and an art, with a substantial learning curve not only for the physician in knowing when to use it but also for the respiratory therapist in tailoring it to the patient's needs and toleration over time. The findings of the present study are very encouraging in their documentation of more widespread use of NIV for AECOPD. Hopefully, the overall improvement in patient outcomes will continue with increasing experience, and the subset of patients who fail initial NIV will be better understood and more successfully managed.


1. Pierson DJ. History and epidemiology of noninvasive ventilation in the acute-care setting. Respir Care 2009;54:40-52.