Failure to accrue subjects. All activity has stopped & no analysis will be done on what has been collected.

Chronic obstructive pulmonary disease (COPD) is a highly prevalent condition worldwide and is a cause of substantial morbidity and mortality. Unfortunately, few therapies have been shown to improve survival. The importance of systemic effects and co-morbidities in COPD has garnered attention based on the observation that many patients with COPD die from causes other than respiratory failure, including a large proportion from cardiovascular causes. Recently, two high profile randomized trials have shown substantial improvements in morbidity and mortality with use of nocturnal non-invasive ventilation (NIV) in COPD patients with hypercapnia. Although the mechanisms by which NIV improves outcomes remain unclear, the important benefits of NIV might be cardiovascular via a number of mechanisms. In contrast to prior trials of NIV in COPD that did not show substantial benefit, a distinguishing feature of these encouraging recent NIV clinical trials was a prominent reduction of hypercapnia, which might be a maker or mediator of effective therapy. Alternatively, improvements might be best achieved by targeting a different physiological measure. Additional mechanistic data are therefore needed to inform future trials and achieve maximal benefit of NIV. Recent work in cardiovascular biomarkers has identified high-sensitivity troponin to have substantial ability to determine cardiovascular stress in a variety of conditions - even with only small changes. In COPD, a number of observational studies have shown that high-sensitivity troponin increases with worsening disease severity, and that levels increase overnight during sleep. This biomarker therefore presents a promising means to study causal pathways regarding the effect of NIV in patients with COPD. With this background, the investigator's overall goals are: 1) To determine whether the beneficial effect of non-invasive ventilation might be due to a reduction in cardiovascular stress, using established cardiovascular biomarkers, and 2) To define whether a reduction in PaCO2 (or alternative mechanism) is associated with such an effect.

Inclusion Criteria:

• Subjects with previously diagnosed severe COPD (FEV1 <50% predicted) and daytime hypercapnia (PaCO2 or TcCO2 > 45 mmHg)

Exclusion Criteria:

• Lung disease besides COPD (e.g., pulmonary fibrosis, bronchiectasis, pulmonary arterial hypertension) other than well controlled asthma
• Unrevascularized coronary artery disease, angina, prior heart attack or stroke, congestive heart failure
• Uncontrolled hypertension (SBP >160, DBP >95)
• Unwilling or unable to withhold CPAP during polysomnography
• Presence of tracheostomy
• Hospitalization within the past 90 days
• Prior peptic ulcer disease, esophageal varicies, or gastrointestinal bleeding (< 5 years)
• Prior gastric bypass surgery
• Anticoagulant use (other than aspirin) or bleeding diathesis (only for esophageal catheter placement)
• Chronic liver disease or end-stage kidney disease
• Allergy to any of the study medications
• Regular use of medications known to affect control of breathing (opioids, benzodiazepines, theophylline)
• Insomnia or circadian rhythm disorder
• Active illicit substance use or >3 oz nightly alcohol use
• Psychiatric disease, other than controlled depression
• Pregnancy
• Prisoners
• Cognitive impairment, unable to provide consent, or unable to carry out research procedures