Recruitment difficulties

Chronic obstructive pulmonary disease (COPD) is a chronic lung disease affecting an estimated 1 in 10 Canadians. Symptoms include persistent shortness of breath, cough and sputum production. The symptoms can be serious when people with COPD experience a flare of their disease and may lead to hospitalization or death. Improving other conditions that affect COPD control is one way to improve the health of people with COPD. Obstructive sleep apnea (OSA) is the most common breathing problem during sleep, and commonly co-exists with COPD. Although diagnosing and treating OSA is encouraged, it has not been highlighted in guidelines that recommend ideal COPD care. People with COPD and OSA have lower sleep quality and lower oxygen levels during sleep compared to people with OSA. Despite these differences, treatment of OSA in people with COPD is modeled after treatment of OSA in the general population, generally using treatment with continuous positive airway pressure (CPAP) with the possible addition of oxygen through the CPAP machine. There are few studies looking at other types of treatment including noninvasive ventilation (NIV) in people with COPD and OSA. The majority of studies of NIV in COPD has been for people with other reasons to use NIV including acute respiratory failure or chronic hypercarbic respiratory failure and did not include people with risk factors for OSA or who had undergone overnight sleep studies.

In Alberta, NIV is provided province wide for people who have both OSA who do not meet certain physiologic targets in their oxygen levels or breathing patterns after CPAP is applied on an overnight sleep study. NIV is provided preferentially to CPAP and oxygen, providing an opportunity to look at health outcomes when NIV is used instead of CPAP for the treatment of patients with COPD. Through this study, we will measure whether people with COPD and a sleep related breathing disorder such as OSA have fewer severe flares of COPD after starting CPAP or NIV. We will evaluate whether the number of Emergency Department visits, hospitalizations or deaths lowers after starting CPAP or NIV.

Chronic obstructive pulmonary disease is a common chronic lung disease affecting more than 10% of Canadians. Patients with COPD have significant morbidity from the disease as almost 60% of Albertans living with COPD were hospitalized between 2012-2014. Sleep related breathing disorders (SRBD) are a common comorbidity of COPD, and contribute to the morbidity associated with COPD. The most common SRBD is OSA. It is estimated that 1.0-3.6% of Canadians are affected by both OSA and COPD. People with COPD and OSA have lower sleep quality, lower oxygen saturations and greater nocturnal oxygen desaturations when their sleep is compared to people with OSA. Observational studies suggest that patients with OSA and COPD have a higher risk of hospitalization, exacerbation of COPD, or death; however, the evidence to support the assessment and management of OSA improving COPD health outcomes has not been considered substantial enough to justify inclusion in COPD guidelines. Other SRBDs that may impact patients with COPD are sleep-related hypoventilation as a consequence of obesity, or central sleep apnea as a consequence of cardiac disease, stroke or use of opioids; however, these are less common and less studied types of SRBD within sleep literature.

Two treatments that are commonly prescribed for SRBD are CPAP or NIV. In patients with OSA, CPAP improves quality of life, daytime sleepiness, and blood pressure control. In patients with COPD who have OSA, observational studies suggest that CPAP use is associated with survival, with the duration of CPAP use per night lowering mortality in a dose dependent manner, an increased time to first moderate-severe COPD exacerbation and fewer hospitalizations related to exacerbations of COPD. The association of decreased health care utilization with CPAP is greater in patients with a higher COPD complexity and number of comorbidities, indicating those who are the sickest may derive the most harm from untreated disease. NIV is better tolerated than CPAP in some patients with OSA; yet, there have been only a few studies evaluating its use for a SRBD in COPD. Further work in this area is critically important as the indications for use of NIV are expanding among patients with COPD.

We are uniquely positioned to complete this study as NIV is provided province-wide in Alberta and patient health outcomes are available in provincial databases. In addition, within the provincial titration protocol for polysomnography, NIV is provided preferentially over CPAP with oxygen, differentiating the management of these patients from the titration protocols used in other studies.

Through this study, we will explore whether treatment with NIV or CPAP for a SRBD in patients with COPD results in fewer moderate-severe exacerbations of COPD and whether the use of these devices have an impact on health care utilization and mortality. In a subgroup analysis, we will explore whether NIV use results in similar outcomes as CPAP, and account for differences in disease severity in this comparison. Adherence with these CPAP or NIV will also be summarized. The results from this study will inform regional care as well as provide insight into the roles of CPAP and NIV in the treatment of SRBD in patients with COPD.

Sample size:

59% of Albertans with COPD experience an exacerbation within a 2 year period; however, a conservative estimate of 50% will be used for the power calculation. A clinically significant difference in COPD exacerbations is estimated to be 25%. An estimated standard deviation of 8 patients hospitalized per 2 years was estimated (variance 64). Given that the exposure group will be stratified by use of CPAP or NIV in subgroup analysis, a 2:1 proportion for the exposure to comparison group was utilized and the exposure arm will be composed of equal proportions of patients prescribed CPAP or NIV. A 10% drop out rate is anticipated.

Primary hypothesis: Null hypothesis (H0): CPAP or NIV is equivalent to no therapy; Alternate hypothesis (H1): CPAP or NIV is better than no therapy

Secondary hypothesis: Null hypothesis (H0): NIV is equivalent to CPAP; Alternate hypothesis (H1): NIV is better than CPAP

For the primary hypothesis, these estimates suggest that a sample size of 165 patients is required (n=110 patients in the exposure arm and 55 patients in the control arm). For the secondary hypothesis, it is not anticipated that a large difference between CPAP or NIV will be observed; however, the subgroup analysis of the primary outcome will be powered to detect a 25% difference between intervention with CPAP or NIV.

Inclusion Criteria:

• Consenting adults who have completed polysomnography at the University of Alberta Sleep Disorders Laboratory in Edmonton, Alberta, Canada.

• Naïve to CPAP or NIV therapy or have use of CPAP or NIV <4 weeks in the past

• Diagnosed with COPD by a respirologist and has fixed airflow obstruction on spirometry (i.e., post-bronchodilator FEV1/FVC <LLN on spirometry or FEV1/FVC <0.70 if the LLN is not provided)

• Diagnosed with a SRBD for which CPAP or NIV are commonly prescribed, as defined below; some of the patients with mild SRBD or a SRBD suspected to be secondary to COPD rather than OSA (e.g. mild OSA, sustained hypercarbia, hypoxemia etc) may not be prescribed CPAP or NIV at the treating physician's discretion

  • OSA will be defined as a predominantly obstructive apnea-hypopnea index (AHI) ≥5 events/hr of sleep on sleep diagnostic testing in the presence of symptoms/relevant comorbidities or an AHI ≥15 events/hr irrespective of symptoms. If home sleep apnea testing was used in place of polysomnography, the respiratory event index (REI) will be a surrogate of AHI.
  • Hypoventilation will be defined by polysomnography criteria as a sustained elevation in the TcCO2 to be >55 mmHg for ≥10 min or a rise in TcCO2 by >10 mmHg to a value >50 mmHg for ≥10 min. If an arterial blood gas is available, sustained hypercapnia as defined by arterial carbon dioxide oxygen tension (PaCO2) >45 mmHg or serum bicarbonate (HCO3) ≥27 mmol/L while medically stable will be reported.

Exclusion Criteria:

• Significant lung disease in addition to COPD
• Neuromuscular disease including diaphragmatic paralysis
• Dementia
• Treatment with CPAP or NIV for >4 weeks prior to study eligibility.
• Refusal to participate
• Inability to consent in English and do not have an interpreter readily available
• Current diagnosis of an active malignancy
• Central sleep apnea as the indication for PAP treatment. Central sleep apnea will be defined as a central hypopnea index ≥5 events/hr of sleep on polysomnography and the number of central apneas and/or central hypopneas is >50% of the AHI with or without Cheyne-Stokes breathing

The study setting are people living in Alberta receiving care by an adult respirologist who undergo sleep diagnostic testing, consent to contact for research at a participating site, are eligible for the study after a brief chart review, and who subsequently consent to be involved in the study. The start date for CPAP or NIV will need to be within the study period although it will not need to be after the date of sleep diagnostic testing in order to allow patients started on NIV in hospital who subsequently undergo sleep diagnostic testing to be eligible. Polysomnograms for eligible patients will be completed by sleep laboratories operating within Alberta Health Services within the public health care system. Home sleep apnea tests are performed by both private and public facilities within Alberta and both sites of testing will be considered adequate for patient inclusion in the study if that is the sole form of sleep diagnostic testing requested by the treating respirologist.

Given that this is a phase IV clinical trial, the information will be disseminated through publication. Requests for IPD data from this study will need to go through institutional ethics review prior to use for a secondary purpose.