Comparative Effectiveness of Noninvasive and Invasive Ventilation in Critically Ill Patients With Acute Exacerbation of Chronic Obstructive Pulmonary Disease

Mihaela S. Stefan, MD; Brian H. Nathanson, PhD, DSHS; Thomas L. Higgins, MD, MBA; Jay S. Steingrub, MD; Tara Lagu, MD, MPH; Michael B. Rothberg, MD, MPH; Peter K. Lindenauer, MD, MSc


Crit Care Med. 2015;43(7):1386-1394. 

In This Article

Abstract and Introduction


Objectives: To compare the characteristics and hospital outcomes of patients with an acute exacerbation of chronic obstructive pulmonary disease treated in the ICU with initial noninvasive ventilation or invasive mechanical ventilation.

Design: Retrospective, multicenter cohort study of prospectively collected data. We used propensity matching to compare the outcomes of patients treated with noninvasive ventilation to those treated with invasive mechanical ventilation. We also assessed predictors for noninvasive ventilation failure.

Setting: Thirty-eight hospitals participating in the Acute Physiology and Chronic Health Evaluation database from 2008 through 2012.

Subjects: A total of 3,520 patients with a diagnosis of chronic obstructive pulmonary disease exacerbation including 27.7% who received noninvasive ventilation and 45.5% who received invasive mechanical ventilation.

Interventions: None.

Measurements and Main Results: Noninvasive ventilation failure was recorded in 13.7% from patients ventilated noninvasively. Hospital mortality was 7.4% for patients treated with noninvasive ventilation; 16.1% for those treated with invasive mechanical ventilation; and 22.5% for those who failed noninvasive ventilation. In the propensity-matched analysis, patients initially treated with noninvasive ventilation had a 41% lower risk of death compared with those treated with invasive mechanical ventilation (relative risk, 0.59; 95% CI, 0.36–0.97). Factors that were independently associated with noninvasive ventilation failure were Simplified Acute Physiology Score II (relative risk = 1.04 per point increase; 95% CI, 1.03–1.04) and the presence of cancer (2.29; 95% CI, 0.96–5.45).

Conclusions: Among critically ill adults with chronic obstructive pulmonary disease exacerbation, the receipt of noninvasive ventilation was associated with a lower risk of in-hospital mortality compared with that of invasive mechanical ventilation; noninvasive ventilation failure was associated with the worst outcomes. These results support the use of noninvasive ventilation as a first-line therapy in appropriately selected critically ill patients with chronic obstructive pulmonary disease while also highlighting the risks associated with noninvasive ventilation failure and the need to be cautious in the face of severe disease.


Chronic obstructive pulmonary disease (COPD) is a highly prevalent condition that is responsible for approximately 1 million hospitalizations each year, and it is the third leading cause of death in the United States.[1–3] There is a wide range of disease severity among hospitalized patients with an acute exacerbation of chronic obstructive pulmonary disease (AE-COPD), ranging from brief hospital admission to prolonged hospitalization and death.[4,5] Approximately 12–18% of patients hospitalized with an AE-COPD are treated in the ICU[6] and mortality in this population approaches 15%.[5]

The efficacy of noninvasive ventilation (NIV) in patients with AE-COPD has been extensively studied. Several randomized controlled trials (RCTs)[7–9] and meta-analyses[10,11] found a reduction in intubation rate, hospital-acquired pneumonia, and mortality when NIV was added to supportive care. A number of guidelines strongly recommend NIV versus standard care alone in moderate to severe COPD exacerbation. However, only two small RCTs directly compared the efficacy of NIV and invasive mechanical ventilation (IMV) and found that NIV use resulted in fewer complications and lower readmission rate without changes in mortality.[12,13] One survey study of 99 patients with AE-COPD admitted to 42 French ICUs[14] and one recent large U.S. study using an administrative dataset[15] showed that NIV use was associated with significant reduction in mortality compared with IMV. Because of insufficient evidence, the Canadian Practice Guidelines and the U.S. Agency for Healthcare Research and Quality comparative effectiveness review make no recommendations about the use of NIV versus IMV in patients with severe acute respiratory failure (ARF) secondary to COPD.[16,17]

There are limited recent data about the use of NIV and its associated outcomes in patients with severe AE-COPD admitted to ICU, and what has been learned recently about the comparative effectiveness of NIV to IMV comes mainly from studies based only on claims data.

We sought to take advantage of a large, multicenter ICU database that contains physiological data to compare the characteristics and short-term outcomes of patients hospitalized with severe COPD exacerbation and treated with NIV and IMV. We hypothesized that after adjusting for severity of illness and other patient and hospital characteristics, patients treated with NIV would have better outcomes than patients treated with IMV.