Positive Airway Pressure Therapy Predicts Lower Mortality and Major Adverse Cardiovascular Events Incidence in Medicare Beneficiaries with Obstructive Sleep Apnea

Diego R Mazzotti, Lemuel R. Waitman, Jennifer Miller, Krishna M. Sundar, Nancy H. Stewart, David Gozal, Xing Song, Greater Plains Collaborative
doi: https://doi.org/10.1101/2023.07.26.23293156

Abstract

Background Obesity is associated with obstructive sleep apnea (OSA) and cardiovascular risk. Positive airway pressure (PAP) is the first line treatment for OSA, but evidence on its beneficial effect on major adverse cardiovascular events (MACE) prevention is limited. Using claims data, the effects of PAP on mortality and incidence of MACE among Medicare beneficiaries with OSA were examined.

Methods A cohort of Medicare beneficiaries with ≥2 distinct OSA claims was defined from multi-state, state-wide, multi-year (2011-2020) Medicare fee-for-service claims data. Evidence of PAP initiation and utilization was based on PAP claims after OSA diagnosis. MACE was defined as a composite of myocardial infarction, heart failure, stroke, or coronary revascularization. Doubly robust Cox proportional hazards models with inverse probability of treatment weights estimated treatment effects controlling for sociodemographic and clinical factors.

Results Among 888,835 beneficiaries with OSA (median age 73 years; 43.9% women; median follow-up 1,141 days), those with evidence of PAP initiation (32.6%) had significantly lower all-cause mortality (HR [95%CI]: 0.53 [0.52-0.54]) and MACE incidence risk (0.90 [0.89-0.91]). Higher quartiles of annual PAP claims were progressively associated with lower mortality (Q2: 0.84 [0.81-0.87], Q3: 0.76 [0.74-0.79], Q4: 0.74 [0.72-0.77]) and MACE incidence risk (Q2: 0.92 [0.89-0.95], Q3: 0.89 [0.86-0.91], Q4: 0.87 [0.85-0.90]).

Conclusion PAP utilization was associated with lower all-cause mortality and MACE incidence among Medicare beneficiaries with OSA. Results might inform trials assessing the importance of OSA therapy towards minimizing cardiovascular risk and mortality in older adults.

Introduction

Obesity is a chronic condition with significant public health burden1, affecting 40% of adults in the U.S.2 and accounting for millions of cardiovascular (CV) disease deaths every year3. There is greater attention about the importance of adequate sleep habits4 and its relation to cardiometabolic risk5,6. Obstructive sleep apnea (OSA) is a highly prevalent disorder contributing to these associations, with population rates of 9%-37% in men and 4-50% in women7, affecting nearly 1 billion people worldwide8. OSA becomes more prevalent with age9,10 and obesity11, and is associated with coronary heart disease, stroke, and CV mortality12,13. The CV risk due to OSA is greater among those with excessive sleepiness14,15, worse nocturnal hypoxemia16, and differential heart rate responses to respiratory events17. Thus, there is increasing attention on therapies that may modify CV risk for primary and secondary prevention5,6,18 through targeting sleep disordered breathing19.

Positive airway pressure (PAP) is the first line of therapy for moderate-severe OSA. Despite epidemiological studies strongly suggesting that OSA is a modifiable CV risk factor20–26, a meta-analysis of randomized controlled trials (RCTs) failed to demonstrate that PAP prevents CV outcomes27. Studies have shown that patient selection and treatment adherence may explain some of the negative results28,29. These challenges are intensified by the lack of large, equitable, and multi-centric pragmatic studies assessing the effectiveness of PAP on improving CV risk. Such studies are expensive and time-consuming, leading to recruitment and retention challenges, and delayed policy changes30,31. Whether long-term PAP therapy prevents CV disease in a clinical population is a critical question that remains unanswered.

Insurance claims data allow the design of observational studies that complement RCTs, by adopting causal inference methods29,32,33. When robustly applied, they have the potential to characterize the effect of OSA therapies in representative clinical settings34. A study on a French nationwide claims database found that continuous PAP (CPAP) therapy termination was associated with all-cause mortality and incidence of heart failure35. In the U.S., Centers for Medicare & Medicaid Services (CMS) beneficiaries represent a large population of older adults with access to healthcare coverage. Studies exploring a 5% sample of randomly selected beneficiaries found that those with OSA had higher healthcare utilization when compared to matched controls36. Moreover, those adherent to CPAP based on Durable Medical Equipment (DME) claims were found to have reduced risk of stroke37 and healthcare expenses among those with pre-existing CV diseases38. However, a major limitation is the reliance only on a sample of beneficiaries, which might represent challenges characterizing the effects of PAP on the incidence of individual CV diseases with accuracy. State-wide Medicare claims, particularly within underserved states in the U.S., might reflect a broader representation of individuals with OSA and provide more generalizable effect estimates.

In this study, we aimed to determine the effect of PAP utilization on all-cause mortality and incidence of MACE and its components among Medicare beneficiaries in the Central U.S. By leveraging a robust causal inference design, we determined the average treatment effect of PAP initiation on the risk of incident MACE and mortality. Next, we established clinically relevant cut-offs of PAP utilization based on DME claims over the first year after PAP initiation to determine PAP utilization exposure groups. We demonstrate that these groups are differentially associated with incident MACE and mortality among older adults. Finally, we provide effect estimates of PAP exposure groups stratified by relevant sociodemographic and clinical factors.

Methods

Additional details are presented in the Supplemental materials.

Study cohort

Medicare beneficiaries (>65 years) enrolled to part A and B, and ≥2 distinct OSA claims were collected from multi-state, state-wide, multi-year (2011-2020) Medicare fee-for-service claims data. State-level Medicare claims data were originally obtained as part of the Greater Plains Collaborative Reusable Observable Unified Study Environment (GROUSE)39, with a catchment area across 11 states in the Central U.S. The study protocol has been approved by Institutional Review Boards at each institution. The study cohort was defined based on a validated EHR algorithm to identify participants with OSA40 (see Supplemental Table 1). We further required >1 year enrolment with Medicare before their first OSA claim, to better capture newly diagnosed OSA and complete PAP utilization history. For the analysis of incident MACE, we further excluded beneficiaries with history of MACE prior to their OSA diagnosis date.

Study design

This is an observational, retrospective analysis of state-wide Medicare claims data. We proposed two complementary causal inference designs to estimate the average treatment effect (ATE) of PAP initiation (Figure 1) and PAP utilization exposure groups (Figure 2). For the PAP initiation analyses, prescription time-distribution matching41 was used to identify the “time zero” (exposure assignment, covariate determination, and start of follow-up) for the group that did not initiate PAP (pseudo landmark date, Figure 1), matched on the distribution of the time from diagnosis and start of PAP therapy in the group that initiated PAP (landmark T, Figure 1). For the analysis of PAP utilization exposure groups, we included beneficiaries that have been enrolled and have not experienced events at the 1-year mark since their PAP initiation date. Exposure groups were defined based on the distribution of claim counts during this first year of PAP utilization and “time zero” was the date of the first anniversary of PAP initiation (Figure 2).

Figure 1.
Figure 1. Overview of the PAP initiation analysis study design.
Figure 2.
Figure 2. Overview of the PAP utilization exposure group study design.

PAP utilization definitions

Evidence of PAP initiation was based on the first PAP initiation claim after OSA diagnosis (Supplemental methods). PAP utilization exposure groups were used as surrogate measures of PAP adherence. We explored different surrogates based on the total count of PAP claims during first year of utilization. Prior studies have suggested that a greater number of DME claims, such as mask refills, correlates with objective adherence based on hours of use42. A comprehensive comparison of all methods is presented in the Supplemental methods and informed the selection of the quartile-based PAP utilization exposure group definition, which was then used in our primary analysis.

Study Outcomes

We assessed two primary outcomes: all-cause mortality and incidence of MACE. All-cause mortality was obtained by identifying date of death provided in the Medicare beneficiary summary file. MACE was defined as a composite of first occurrence of myocardial infarction (MI), heart failure (HF), stroke or coronary revascularization, identified by diagnostic and procedure codes (Supplemental Table 1). Secondary analyses using each component of MACE as outcomes are also presented.

Covariates

We included the following covariates: age, sex, race (White, Black, American Indian, Asian, Other), socioeconomic status, prior history of type 2 diabetes, hypertension, obesity, atrial fibrillation, MACE (only in models assessing all-cause mortality), chronic obstructive pulmonary disease (COPD), chronic kidney disease (CKD), anxiety disorder, hypersomnia, insomnia, Charlson comorbidity index (CCI)43,44, prescriptions of anticoagulants, antihypertensives, antilipidemic agents and blood glucose regulators (see Supplemental Table 1 for definitions).

Statistical analyses

Sociodemographic characteristics and clinical history were described between PAP initiation exposure groups using counts and percentages for categorical data and median and interquartile range for continuous data. Univariate associations between sociodemographic and clinical history variables with PAP initiation exposure groups were performed using chi-squared tests. We used Kaplan-Meier survival analyses and log-rank tests to compare survival curves between PAP initiation exposure groups for MACE and all-cause mortality. We assessed the proportional hazards assumption by assessing scaled Schoenfeld residuals, which did not find strong evidence of violation. We implemented a causal inference framework to determine the effect of PAP initiation or PAP utilization exposure groups on the outcomes of interest (all-cause mortality and MACE. Specific details are presented in the Supplemental methods. Briefly, we derived propensity scores (PS) based on PAP initiation or utilization, and then calculated weights that were used in fully adjusted weighted Cox regression models assessing the effect of PAP initiation or PAP utilization exposure groups on the outcomes of interest, representing causal estimate of the ATE derived from a doubly robust estimator. This framework was applied to our primary analyses, as well as to secondary analyses assessing individual components of MACE. Results of stratified analysis by sociodemographic and clinical characteristics are also presented. We determined statistical significance based on Bonferroni-corrected thresholds of p<0.025 (2 primary outcomes) and provide E-values to assess the strength of association a potential unmeasured confounder would need to express to nullify observed associations45.

Results

Sample characterization

Our sample included 888,835 eligible Medicare beneficiaries with evidence of OSA diagnosis (median [Q1, Q3] age 73 [69, 78] years; 43.9% women; median [Q1, Q3] follow-up 3 [1.5, 5.1] years). Figure 3 shows the study flowchart. Among eligible participants, 290,015 (32.6%) had evidence of PAP initiation. Table 1 shows the sample characteristics according to PAP initiation groups. Participants that have initiated PAP were younger, more likely to be women and White, more likely to have been diagnosed with hypersomnia and insomnia, and less likely to be diagnosed with COPD, type 2 diabetes, hypertension, obesity, atrial fibrillation, CKD, anxiety disorders, and to have had a first MACE at OSA diagnosis. Overall, 5-year cumulative mortality rate was 20.5% (based on total N=888,835) and 5-year cumulative MACE incidence was 41.0% (based on those without MACE at baseline, N=572,072). When stratified by PAP initiation exposure groups, those who initiated PAP had a 5-year cumulative mortality rate and cumulative MACE incidence of 12.4% and 27.4% respectively. Those who did not initiate PAP had a 5-year incidence rates of mortality and MACE event of 17.7% and 30.7%, respectively (p<0.001 for both).

Figure 3.
Figure 3. Study cohort definitions and CONSORT diagram.
View this table:
Table 1: Sample characteristics by positive airway pressure initiation groups among eligible Medicare beneficiaries.

PAP initiation is associated with greater survival and MACE-free probability

Kaplan-Meier survival curves for all-cause mortality and MACE, according to PAP initiation exposure groups are presented in Figure 4. Log-rank test indicates significant differences in all-cause mortality and MACE-free survival probabilities (p<0.001), with participants that have initiated PAP presenting greater survival than those without.

Figure 4:
Figure 4: Unadjusted and adjusted Kaplan-Meier survival curves describing the survival probabilities between PAP initiation exposure groups and all-cause mortality (top) and MACE (bottom). Numbers in boxes represent annual survival probabilities per group. The results of log-rank tests comparing exposure groups are also shown.

PAP initiation is independently associated with lower mortality and lower incidence of MACE

Results of doubly robust Cox proportional hazards models are presented in Figure 5. Patients with evidence of PAP initiation had significantly lower all-cause mortality risk (HR [95%CI] = 0.53 [0.52-0.54]; p<0.001) and lower MACE incidence risk (0.90 [0.89-0.91]; p<0.001) when compared to those without evidence of initiating PAP. The risk ratio of an unmeasured confounder (E-value) would need to be 2.47 to explain away the effect of PAP initiation on mortality, and 1.36 to explain away the effect on MACE.

Figure 5.
Figure 5. Summary of IPTW-adjusted Cox proportional hazards models assessing the effect of PAP initiation on all-cause mortality (ACM) and MACE in all eligible participants. Results were derived from IPTW-weighted Cox proportional hazards models adjusted for age, sex, race, low-income-subsidy or dual-eligibility indicator, type 2 diabetes, hypertension, obesity, atrial fibrillation, MACE (all-cause mortality only), COPD, CKD, hypersomnia, and insomnia, anxiety disorder, hypersomnia, insomnia, CCI, prescriptions of anticoagulants, antihypertensives, antilipidemic agents and blood glucose regulators. Reference category is no evidence of PAP initiation. Abbreviations: ACM: all-cause mortality; COPD: chronic obstructive pulmonary disease; CKD: chronic kidney disease; CCI: Charlson comorbidity index; PAP: positive airway pressure; HR: hazard ratio; CI: confidence interval; MACE: major adverse cardiovascular event; IPTW: inverse probability of treatment weights. and within categories of relevant sociodemographic and clinical characteristics. Among stratified models, stratification variable was not included as a covariate in propensity score models nor in outcome models.

Analyses of secondary outcomes (MI, HF, stroke, and coronary revascularization) are presented in Figure 6. OSA patients with evidence of PAP initiation had significantly lower incidence risk of MI (0.84 [0.82-0.85]; p<0.001; E-value: 1.48), HF (0.89 [0.88-0.90]; p<0.001; E-value: 1.39), and stroke (0.86 [0.84-0.88]; p<0.001; E-value: 1.41). Moreover, PAP initiation was associated with higher incidence risk of revascularization (1.08 [1.05-1.11]; p<0.001).

Figure 6.
Figure 6. Summary of IPTW-adjusted Cox proportional hazards models assessing the effect of PAP initiation on myocardial infarction (MI), heart failure (HF), stroke (STROKE) and coronary revascularization (REVASC) in all eligible participants. Results were derived from IPTW-weighted Cox proportional hazards models adjusted for age, sex, race, low-income-subsidy or dual-eligibility indicator, type 2 diabetes, hypertension, obesity, atrial fibrillation, COPD, CKD, hypersomnia, and insomnia, anxiety disorder, hypersomnia, insomnia, CCI, prescriptions of anticoagulants, antihypertensives, antilipidemic agents and blood glucose regulators. Reference category is no evidence of PAP initiation. Abbreviations: ACM: all-cause mortality; COPD: chronic obstructive pulmonary disease; CKD: chronic kidney disease; CCI: Charlson comorbidity index; PAP: positive airway pressure; HR: hazard ratio; CI: confidence interval; MACE: major adverse cardiovascular event; IPTW: inverse probability of treatment weights.

Next, we investigated whether the effect of PAP initiation was consistent across subgroups of sociodemographic and clinical factors. Results of primary outcomes are presented in Supplemental Figure 1 and of secondary outcomes are presented in Supplemental Figure 2. We observed that effect estimates regarding all-cause mortality are consistent across subgroups, with significantly lower risk among those that initiated PAP across all strata, and key differences in certain subgroups. We observed stronger protective effects of PAP initiation among women (0.52 [0.51-0.53]) as compared to men (0.53 [0.53-0.54), those with evidence of obesity diagnosis (0.51 [0.50-0.51]) as compared to those without (0.55 [0.54-0.55]), those with evidence of history of atrial fibrillation (0.47 [0.46-0.48]) as compared to those without (0.54 [0.53-0.54]), and those without evidence of hypersomnia (0.52 [0.52-0.53]) as compared to those with (0.63 [0.61-0.65).

Regarding incidence of MACE, we observed stronger PAP initiation effects among those 75 years or older (75-79 years: 0.89 [0.87-0.90]; 80+: 0.86 [0.85-0.88], as compared to those <75 years (65-69 years: 0.91 [0.90-0.93]; 70-74 years: 0.92 [0.90-0.93]). We also observed stronger protective effects of PAP initiation among women (0.88 [0.87-0.89]) as compared to men (0.91 [0.90-0.92]), those with lower social economic status (0.86 [0.85-0.87]) as compared to those with higher (0.91 [0.91-0.92]), those with evidence of insomnia diagnosis (0.87 [0.85-0.89]) as compared to those without (0.90 [0.89-0.91]), those with evidence of obesity diagnosis (0.88 [0.87-0.89]) as compared to those without (0.91 [0.90-0.92]), those with evidence of COPD diagnosis (0.88 [0.86-0.89]) as compared to those without (0.90 [0.90-0.91]), those with evidence of type 2 diabetes diagnosis (0.88 [0.87-0.89]) as compared to those without (0.91 [0.90-0.92]), those with evidence of atrial fibrillation diagnosis (0.85 [0.83-0.87]) as compared to those without (0.90 [0.90-0.91]), and those with higher CCI scores (CCI=1-2 comorbidities: 0.88 [0.87-0.90]; CCI=3-4 comorbidities: 0.89 [0.88-0.91]; CCI=5+ comorbidities: 0.89 [0.87-0.91]) as compared to those with less comorbidities (CCI=0: 0.94 [0.92-0.96]).

Stratified analyses regarding secondary outcomes (MACE components) consistently showed significantly lower incidence of MI, HF, and stroke as well as higher incidence of coronary revascularization procedures across all subgroups of sociodemographic and clinical factors (Supplemental Figure 2). Important differences were observed for certain subgroups. Regarding MI, strongest beneficial effects of PAP initiation were observed among women (0.80 [0.79-0.82]), those with lower social economic status (0.80 [0.78-0.82]), those with evidence of hypersomnia diagnosis (0.77 [0.73-0.81]), those with evidence of insomnia diagnosis (0.79 [0.76-0.82]), and those with evidence of anxiety disorder diagnosis (0.79 [0.76-0.82]). Regarding HF, strongest beneficial effects of PAP initiation were observed in those 75 years or older (75-79 years: 0.86 [0.84-0.88]; 80+ years: 0.86 [0.84-0.88]), and those with evidence of atrial fibrillation diagnosis (0.84 [0.81-0.87]). Regarding stroke, strongest beneficial effects of PAP initiation were observed among women (0.83 [0.81-0.85]), those with lower social economic status (0.82 [0.80-0.84]), those with evidence of obesity diagnosis (0.82 [0.80-0.84]), those with evidence of CPOD diagnosis (0.81 [0.79-0.84]), those with evidence of hypertension diagnosis (0.85 [0.83-0.86]), and those with highest CCI scores (0.82 [0.79-0.84]). Regarding coronary revascularization, strongest detrimental effects of PAP initiation were observed among those with 80+ years (1.18 [1.12-1.24]), men (1.10 [1.08-1.13]), and those taking antihypertensive medications (1.12 [1.09-1.15]).

Determination of first year PAP utilization exposure groups

To inform our proposed analysis assessing the effect of PAP utilization exposure groups, we first observed the distribution of total PAP claims per patient at the end of first year of PAP utilization (Supplemental Figure 3). PAP claim counts showed a bimodal distribution, suggesting two utilization patterns were more prevalent: a mode on 3 claims and a mode on 14 claims, which may be suggestive of CMS PAP reimbursement models46 and may correlate with PAP adherence47,48. Because Medicare claims databases do not include PAP adherence tracking, we explored different PAP exposure group definitions (see PAP utilization definitions and Supplemental Figure 4) to determine relevant cut-offs in the distribution of PAP claims. We further assessed associations between different PAP exposure group definitions and our primary and secondary outcomes. In general, a greater count of PAP claims during the first year was progressively associated with lower all-cause mortality and lower MACE incidence risk (Supplemental Figure 5), reaching a plateau after approximately 15-16 claims. When using the numeric count of PAP claims as our exposure, each additional PAP claim was significantly associated with lower hazards of all-cause mortality (0.98 [0.98-0.98]) and MACE (0.99 [0.99-0.99]). To address the non-linearity of the associations between PAP claim counts and outcomes, we performed spline extrapolation analyses and estimated the HR as a function of PAP claim counts (Supplemental Figure 6). Results suggest HR estimates are the highest (e.g., greater hazards) around 4-5 PAP claims and lowest at 17-19 PAP claims across all study outcomes. Taken together, these exploratory analyses helped to inform us that using a quartile-based definition of PAP exposure groups based on claim counts (Q1: 1 to 7 claims; Q2: 8 to 12 claims; Q3: 13 to 15 claims; and Q4: >15 claims) provided a realistic representation of PAP utilization patterns that demonstrated clinically relevant variation in incident outcome risk. Therefore, we focused the description of the following results based on this definition.

Greater PAP utilization is progressively associated with lower risk of mortality and MACE

Results of doubly robust Cox proportional hazards models assessing the effect of PAP utilization exposure groups based on quartiles of PAP claim counts on all-cause mortality and incidence of MACE are presented in Figure 7 (among all participants) and Supplemental Figure 7 (stratified by sociodemographic and clinical characteristics), and results for secondary outcomes are presented in Figure 8 and Supplemental Figure 8. Among patients with OSA and with evidence of PAP utilization in their first year of therapy, higher quartiles of claim counts (i.e., higher PAP utilization) were progressively associated with lower all-cause mortality (Q2: 0.84 [0.81-0.87], Q3: 0.76 [0.74-0.79], Q4: 0.74 [0.72-0.77]) and lower MACE incidence (Q2: 0.92 [0.89-0.95], Q3: 0.89 [0.86-0.91], Q4: 0.87 [0.85-0.90]), when compared to those with ≤8 PAP claims in their first year. Similar results were observed for secondary outcomes, with lower MI incidence risk on Q3 (0.88 [0.83-0.93]) and Q4 (0.84 [0.79-0.88]), lower HF incidence risk (Q2: 0.91 [0.88-0.95], Q3: 0.86 [0.83-0.89], Q4: 0.85 [0.82-0.88]) and lower stroke incidence risk (Q2: 0.85 [0.80-0.91], Q3: 0.86 [0.81-0.90], Q4: 0.80 [0.76-0.85]). No evidence of association with incident coronary revascularization was observed (Figure 8).

Figure 7.
Figure 7. Summary of IPTW-adjusted Cox proportional hazards models assessing the effect of PAP utilization exposure groups based on quartiles of PAP claim counts on all-cause mortality (ACM) and MACE among all participants. PAP exposure groups were defined as follows: Q1 (reference): 1-7 claims; level 2: 8-12 claims; level 3: 13-15 claims; level 4: >15 claims). Results were derived from IPTW-weighted Cox proportional hazards models adjusted for age, sex, race, low-income-subsidy or dual-eligibility indicator, type 2 diabetes, hypertension, obesity, atrial fibrillation, MACE (all-cause mortality only), COPD, CKD, hypersomnia, and insomnia, anxiety disorder, hypersomnia, insomnia, CCI, prescriptions of anticoagulants, antihypertensives, antilipidemic agents and blood glucose regulators. Abbreviations: ACM: all-cause mortality; COPD: chronic obstructive pulmonary disease; CKD: chronic kidney disease; CCI: Charlson comorbidity index; PAP: positive airway pressure; HR: hazard ratio; CI: confidence interval; MACE: major adverse cardiovascular event; IPTW: inverse probability of treatment weights.
Figure 8.
Figure 8. Summary of IPTW-adjusted Cox proportional hazards models assessing the effect of PAP utilization exposure groups based on quartiles of PAP claim counts on myocardial infarction (MI), heart failure (HF), stroke (STROKE) and coronary revascularization (REVASC) in all eligible participants. PAP exposure groups were defined as follows: Q1 (reference): 1-7 claims; level 2: 8-12 claims; level 3: 13-15 claims; level 4: >15 claims). Results were derived from IPTW-weighted Cox proportional hazards models adjusted for age, sex, race, low-income-subsidy or dual-eligibility indicator, type 2 diabetes, hypertension, obesity, atrial fibrillation, COPD, CKD, hypersomnia, and insomnia, anxiety disorder, hypersomnia, insomnia, CCI, prescriptions of anticoagulants, antihypertensives, antilipidemic agents and blood glucose regulators. Abbreviations: ACM: all-cause mortality; COPD: chronic obstructive pulmonary disease; CKD: chronic kidney disease; CCI: Charlson comorbidity index; PAP: positive airway pressure; HR: hazard ratio; CI: confidence interval; IPTW: inverse probability of treatment weights.

Analyses stratified by sociodemographic and clinical characteristics (Supplemental Figure 7) were consistent overall, supporting the protective effects of higher PAP claim counts against all-cause mortality and incident MACE across several strata. Exceptions include certain race groups, likely due to the limited sample size of subgroups resulting in wide confidence intervals and less accurate estimates. Regarding all-cause mortality, key differences across strata include stronger protective effects of higher PAP utilization among those with 65-69 years, and those without evidence of hypertension. Regarding MACE, stronger protective effects of higher PAP utilization were observed among those with 65-69 years, those with higher socioeconomic status, those without history of hypertension, those without evidence of atrial fibrillation, and those without evidence of insomnia. Regarding secondary outcomes, (Supplemental Figure 8), stronger protective effects of higher PAP utilization were observed against MI among those with 80+ years, those with lower socioeconomic status, and those with evidence of hypersomnia. Regarding HF, stronger protective effects of higher PAP utilization were observed among those with 65-69 years, those without evidence of insomnia, those without evidence of hypertension, and those within less comorbidities according to the CCI. Regarding stroke, protective effects of higher PAP utilization were observed among those without evidence of insomnia. Finally, no consistent associations were observed regarding coronary revascularization procedures. Effects estimates and confidence intervals are shown Supplemental Figures 7 and 8.

Discussion

This study provided a systematic assessment of the effects of PAP initiation and utilization on MACE and mortality among older adults that are Medicare beneficiaries across 11 states in the U.S. Our main findings suggest significant associations between claims-based PAP initiation and utilization and lower all-cause mortality and incidence of MACE. Associations of PAP initiation and utilization were also significant for MACE components, with consistent effects on MI, HF, and stroke. We observed that PAP initiation effects were stronger in women (for both mortality and MACE) in those 75 years or older (MACE only), with lower socioeconomic status (MACE only) and with clinical comorbidities (both mortality and MACE). Higher PAP utilization exposure groups were progressively associated with lower incidence of primary and secondary outcomes, with consistent effects in stratified analysis, but with particularly stronger beneficial effects among those between 65-69 years and those without evidence of hypertension (both mortality and MACE), as well as those with higher socioeconomic status, those without evidence of atrial fibrillation, and those without evidence of insomnia (MACE only). We also found that PAP initiation was associated with higher incidence risk of revascularization, with stronger effects among those 80+ years, men, and those taking antihypertensive medications. We used a robust causal inference framework to minimize immortal time bias and provided more reliable estimates of treatment effects. Results from this observational study might inform future trials assessing the importance of OSA therapy initiation and maintenance towards minimizing CV risk and mortality in older adults.

Extensive epidemiological evidence suggests that moderate-severe OSA is associated with hypertension49–51, stroke26, atrial fibrillation52 MI53, and CV mortality54. Several mechanisms may explain the relationship between OSA and CV diseases, including sympathetic activation, endothelial dysfunction, oxidative stress, systemic inflammation, cardiac remodeling due to hypoxemia, intrathoracic pressure swings and changes in blood pressure55. These overlapping mechanisms indicate that therapies targeting OSA are likely to be beneficial towards CV disease prevention. RCTs have established that PAP improves daytime symptoms, mood, and quality of life56,57, and reduces systolic and diastolic blood pressure49, especially in patients with resistant hypertension51. PAP therapy has positive effects on long-term survival in patients with ischemic stroke and moderate-severe OSA58 and it has been associated with lower rate of fatal and non-fatal CV events over a 10-year follow-up53.

Despite this consistent evidence, recent RCTs59–61 investigating the effect of CPAP on different fatal and non-fatal CV events were negative27. These findings were supported by a recent individual meta-analysis of these studies62, although a post-hoc on-treatment analysis suggested a protective effect of good adherence to CPAP and reduced risk of MACE62. A careful investigation of the limitations in these trials29 has suggested that patient selection and treatment adherence may have played an important role on explaining differences between RCTs and observational studies. In fact, less than 20% of patients seeking care in sleep clinics would be eligible to participate in these RCTs63. Many patients present with excessive daytime sleepiness, which was an important exclusion criterion in these trials, due to the ethical reason of randomizing patients that are excessively sleepy into control arms29. However, we14,64 and others15,65 have demonstrated this subgroup is at increased risk for incident CV diseases. These observations suggest that alternative approaches using causal inference in well-designed observational studies could provide more generalizable evidence towards the role of PAP therapy on CV outcomes risk.

Exploring ‘real-world data’ approaches using robust statistical methods that minimize biases and confounding issues that are inherent to observational studies may help inform evidence generation in a more timely and cost-effective manner. The success of these approaches has been demonstrated by the current investigation, as it provided relevant and generalizable evidence about the beneficial effects of PAP initiation and utilization in a defined population of older adults in the U.S.. It has also been demonstrated by recent studies exploring the role of CPAP on risk of mortality, HF35, stroke37, reduced health-care utilisation66 and related expenses among those with pre-existing CV diseases38, including HF with reduced ejection fraction67. In the context of a recently published report from the AHRQ suggesting a lack of strong evidence supporting the role of PAP on long-term CV outcomes68, these studies are fundamental to fill this evidence gap.

Our study has identified important associations within subgroups. Notably, we observed greater beneficial effects of PAP initiation on MACE among older adults ≥75 years, suggesting that primary CV prevention among those with newly diagnosed OSA at older ages could be beneficial. However, these results are conflicting with a 3-month RCT of CPAP therapy among elderly adults >70 years with moderate-severe OSA, which did not find significant effects on blood pressure levels or neurocognitive tests, although significant improvements in sleepiness were observed69. A secondary analysis of this trial among those >80 years did not find significant effects on sleepiness70. Among those that initiated PAP therapy, our study found greater beneficial effects of PAP utilization in those that are 65-69 years. Differences may be explained by length of follow-up and study design considerations. Future trials assessing the role of PAP therapy on health outcomes of older adults using longer follow-up times are warranted. We also observed stronger PAP initiation and greater utilization effects on mortality and MACE among women, in agreement with prior studies, particularly post-menopause56. It is likely that women are subjected to stronger acute detrimental effects of OSA following menopause, and therefore might experience stronger therapeutic effects. Women with OSA are also more likely to have comorbid metabolic conditions71. However, little is known about sex differences in PAP treatment responses72, particularly at older ages. Patients with lower socioeconomic status may be less receptive to receiving PAP therapy73 and report poorer treatment adherence74. Yet, our results demonstrate that those with low-income subsidy and dual Medicare/Medicaid eligibility have greater benefit of PAP against MACE. These results highlight opportunities to implement care programs designed to minimize health disparities, as recently proposed75. Our study also identified stronger effects of PAP initiation among those with comorbid conditions, particularly obesity and atrial fibrillation (for both mortality and MACE), as well as COPD, type 2 diabetes, and those with more comorbidities according to the CCI. It is likely that these subgroups are at a greater underlying risk with the added comorbid impact12,76–78, and therefore therapy benefit might be greater. Importantly, subgroups with evidence for comorbidities might also represent patients at highest risk for early termination of CPAP79, which suggests that efforts to support therapy continuation might be beneficial particularly for these patients.

We also found important differences among those with prior history of hypersomnia or insomnia. While we found stronger beneficial effects of PAP initiation against mortality among those without evidence of hypersomnia, we observed stronger effects of PAP initiation against MACE among those with evidence of hypersomnia, with results being driven by stronger effects against MI. Results were also corroborated by stronger effects of greater PAP utilization against MI among those with evidence of hypersomnia. These results support the well-established epidemiological relationship between excessive daytime sleepiness and increased CV risk among those with OSA14,15,64,65. It is important to note that granular measures of excessive daytime sleepiness might provide a more robust phenotype when compared to claims for hypersomnia. Moreover, only ∼9% of our study cohort had evidence of hypersomnia, which suggests it might not capture all patients with the excessively sleepy phenotype. We also observed stronger effects of PAP initiation against MACE among those with evidence of insomnia. These results also support studies that demonstrate greater CV risk associated with comorbid insomnia and sleep apnea (COMISA). In summary, these results might help guide future RCTs assessing the role of PAP initiation and/or strategies to increase PAP adherence among higher-risk subgroups.

Unexpected results between PAP initiation and increased risk of coronary revascularization procedures, suggesting potentially damaging effects of PAP. Interestingly, these results were mostly driven by older beneficiaries (>75 years). Moreover, analysis of PAP utilization based on claims did not show significant effects of greater PAP utilization and increased risk of revascularization. Future studies assessing the role of PAP initiation on the incident risk of revascularization procedures might provide further insights.

This study utilized state-wide Medicare claims data to determine the effects of PAP initiation and utilization on outcomes. While we acknowledge that objective definitions of PAP adherence and efficacy are preferred to estimate exposure to PAP, telemonitoring data has only become available in recent years80. Prior studies have investigated similar PAP exposure definitions in a nationwide claims database in France35,79 and within 5% Medicare beneficiaries samples in the U.S.36–38,81, and reported consistent results as with objective PAP adherence data66,67. From an epidemiologic perspective, state-wide claims databases cover a large proportion of the older adult demographic and offer a scalable approach to identify those likely to be exposed to PAP independently of vendor or telemonitoring capabilities, thus maximizing generalizability. A single-site study comparing objective adherence based on hours of use found that mask refill claims positively correlated with PAP utilization42. To provide further information about PAP utilization based on claims, our study presented a detailed assessment of different PAP utilization exposure group definitions. We observed a bimodal distribution of PAP claim counts within the first year since PAP initiation, which is suggestive of recognized CMS reimbursement models 46. CMS reimburses PAP therapy among eligible beneficiaries in monthly instalments; when a patient does not meet the eligibility criteria for continued therapy at the third month of utilization, the device is not reimbursed, and no further claims will be identified. We speculate that the first mode of the distribution (3 claims) may represent patients that did not meet the CMS criteria for adherence and therefore may present lower overall PAP utilization during their first year. Similarly, the second mode (14 claims) may represent patients that have demonstrated early adherence and continued to use PAP throughout the year. Our results are consistent with this definition, as it suggests that greater number of PAP claims are progressively associated with greater beneficial outcomes. Therefore, while large scale multisite studies predicting short-and long-term objective PAP adherence using claims data are still needed, the definitions used in our study seem to capture relevant clinical outcomes regarding PAP utilization.

Our study has important limitations. First, our cohort identification approach was based on ICD codes, available through a Medicare claims database, thus preventing characterization of disease severity based on the apnea-hypopnea index (AHI) or other novel physiological traits that have been recently demonstrated to be important for CV risk stratification, such as hypoxic burden16 and heart rate responses to events17. However, we used an algorithm validated in different clinical sites that demonstrated excellent predictive performance40, suggesting that our cohort definition reliably captured individuals with OSA. Second, our claims-based PAP utilization definition may not accurately represent more objective definitions of PAP use. However, hours of use (often used as the only exposure when telemonitoring data is available) may not be the only determinant of PAP efficacy, as other factors such as proportion of protected sleep time (e.g., how many hours sleep on efficacious PAP relative to total sleep time), efficacy of therapy reflected by the optimal pressure, minimal leak, and minimal residual AHI, and even timing of PAP utilization during the night (e.g., early morning termination of therapy might induce more severe rapid eye movement sleep AHI reducing overall PAP efficacy82,83), might be more important predictors of successful therapy, and are yet to be consistently assessed in telemonitoring studies. We encourage that PAP vendors expand academic-industry partnerships towards facilitating privacy-preserving linkage of PAP telemonitoring databases with more granular information from EHR, claims, patient reported outcomes and wearables. Such approaches will be fundamental in future studies aimed at determining the effects of PAP therapy on both short-and long-term clinical outcomes. Third, our covariate definition was based only on claims data, and lack granular information about other relevant confounders such as body mass index, blood pressure, and laboratory measurements that are predictive of CV risk. Moreover, other important factors such as diet, physical activity, and healthy adherer behaviours84 were not available and could explain some of the observed effects. To mitigate this limitation, we provide E-values as a guide to understand the impact of a potential unmeasured confounders in our reported associations. Results suggest that unmeasured confounders with effects (in the risk ratio scale) as large as 2.47 for all-cause mortality and 1.36 for MACE would have to be observed, after adjustment for other covariates, to explain away associations with our primary outcomes. Finally, the present study focuses on a population of older adults that are Medicare beneficiaries in the Central U.S., and extrapolation of effect estimates to other demographics cannot be made. In one hand, studies in mid-life adults would provide important insights about OSA natural history and early prevention of CV outcomes. On the other hand, our study is generalizable to a very large proportion of older adults, as >95% of those >65 years in the United States are enrolled in Medicare. There is also greater confidence in the ascertainment of outcomes.

In conclusion, PAP utilization based on claims was associated with lower all-cause mortality and MACE incidence (including MI, HF, and stroke) among Medicare beneficiaries >65 years. Results support the hypothesis that PAP has beneficial effects against mortality and CV diseases. This study has the potential to inform future trials assessing the importance of OSA therapy initiation and maintenance towards minimizing adverse health outcomes in older adults leading to longer and healthier lives. These results may also inform strategies to improve adherence and efficacy of PAP therapy there are patient-specific, towards personalized sleep medicine.

Funding

The study was funded by the American Heart Association (20CDA35310360), Patient-Centered Outcomes Research Institute (RI-CRN-2020-003-IC); NIH CTSA NCATS Frontiers: University of Kansas Clinical and Translational Science Institute (UL1TR002366); Tier 2 grant, University of Missouri.

Conflict of interest statement

Dr. Krishna Sundar was a past advisory board member for Resmed Inc. and is the co-founder of Hypnoscure LLC through the University of Utah Technology Commercialization Office All other authors declared no conflicts of interest.

Data availability statement

The data underlying this article cannot be shared publicly to protect the privacy of individuals that participated in the study, as stipulated by each institution of the Greater Plains Collaborative. There are established processes for interested investigators to collaborate with the Greater Plains Collaborative research network, that can be accessed via this URL: https://gpcnetwork.org/.

Notes

Greater Plains Collaborative

Sravani Chandaka, Kelechi (KayCee) Anuforo, Lav Patel, Daryl Budine, Nathan Hensel, Siddharth Satyakam, Sharla Smith, Dennis Ridenour, Cheryl Jernigan, Carol Early, Kyle Stephens, Kathy Jurius, Abbey Sidebottom, Cassandra Rodgers, Hong Zhong, Vino Raj, Victor Melendez, Angie Hare, Roman Melamed, Curtis Anderson, Thomas Schouweile, Christine Roering, Philip Payne, Snehil Gupta, John Newland, Albert Lai, Joyce Balls-Berry, Janine Parham, Evin Fritschle, Shanelle Cripps, Kirk Knowlton, Channing Hansen, Erna Serezlic, Benjamin Horne, Jeff VanWormer, Judith Hase, Janet Southworth, Eric Larose, Mary Davis, Laurel Hoeth, Sandy Strey, Brad Taylor, Kris Osinski, April Haverty, Alex Stoddard, Sarah Cornell, Phoenix Do, Lucy Bailey, Beth McDonough, Betsy Chrischilles, Ryan Carnahan, Brian Gryzlak, Gi-Yung Ryu, Katrina Oaklander, Pastor Bruce Hanson, Brad McDowell, Jarrod Field, Abu Mosa, Sasha Lawson, Jim McClay, Soliman Islam, Vasanthi Mandhadi, Kim Kimminau, Dennis Ridenour, Jeff Ordway, Bill Stephens, Russ Waitman, Deandra Cassone, Xiaofan Niu, Lisa Royse, Vyshnavi Paka, Lori Wilcox, Janelle Greening, Carol Geary, Goutham Viswanathan, Jim Svoboda, Jim Campbel, Frances (Annette) Wolfe, Haddy Bah, Todd Bjorklund, Jackson Barlocker, Josh Spuh, Louisa Stark, Mike Strong, Otolose Fahina Tavake-Pas, Rachel Hess, Jacob Kean, Sarah Mumford, Ainsley Huffman, Annie Risenmay, Olivia Ellsmore, Lissa Persson, Kayla Torres Morales, Sandi Stanford, Mahanaz Syed, Rae Schofield, Meredith Zozus, Brian Shukwit, Matthew Decaro, Natalia Heredia, Charles Miller, Alice Robinson, Elmer Bernstam, Fatima Ashraf, Shiby Antony, Juliet Fong Zechner, Philip Reeder, Cindy Kao, Kate Wilkinson, Tracy Greer, Alice Robinson, Lindsay Cowell

Acknowledgements

The authors would like to acknowledge the technical and scientific resources and support from the Greater Plains Collaborative team members, as well as all Medicare beneficiaries whose data were contributing to this investigation.

Footnotes

  • The study incorporated additional Medicare claims data from 2018-2020, substantially increasing the sample size. As a consequence, all reported estimates have been updated, including figures and supplemental materials. The conclusions were very similar with some additional results that are discussed in greater detailed in this version.

Abbreviations list

ATE
average treatment effect
CCI
Charlson comorbidity index
CI
confidence interval
CMS
Centers for Medicare & Medicaid Services
COPD
chronic obstructive pulmonary disease
CPAP
continuous positive airway pressure
CV
cardiovascular
DME
durable medical equipment
HF
heart failure
HR
hazards ratio
MACE
major adverse cardiovascular events
MI
myocardial infarction
OSA
obstructive sleep apnea
PAP
positive airway pressure
PS
propensity scores
RCT
randomized controlled trials.

References

  1. 1.↵
    Powell-Wiley TM, Poirier P, Burke LE, et al. Obesity and Cardiovascular Disease: A Scientific Statement From the American Heart Association. Circulation. 2021;143(21).
  2. 2.↵
    Flegal KM, Kruszon-Moran D, Carroll MD, Fryar CD, Ogden CL. Trends in Obesity Among Adults in the United States, 2005 to 2014. Jama. 2016;315(21).
  3. 3.↵
    The GBD 2015 Obesity Collaborators. Health Effects of Overweight and Obesity in 195 Countries over 25 Years. New England Journal of Medicine. 2017;377(1):13–27.
    OpenUrlCrossRefPubMed
  4. 4.↵
    Hall WL. The emerging importance of tackling sleep–diet interactions in lifestyle interventions for weight management. British Journal of Nutrition. 2022;128(3):561–568.
    OpenUrlCrossRef
  5. 5.↵
    Tsao CW, Aday AW, Almarzooq ZI, et al. Heart Disease and Stroke Statistics-2022 Update: A Report From the American Heart Association. Circulation. 2022;145(8):e153–e639.
    OpenUrlCrossRefPubMed
  6. 6.↵
    Lloyd-Jones DM, Allen NB, Anderson CAM, et al. Life’s Essential 8: Updating and Enhancing the American Heart Association’s Construct of Cardiovascular Health: A Presidential Advisory From the American Heart Association. Circulation. 2022;146(5).
  7. 7.↵
    Franklin KA, Lindberg E. Obstructive sleep apnea is a common disorder in the population-a review on the epidemiology of sleep apnea. J Thorac Dis. 2015;7(8):1311–1322.
    OpenUrlCrossRefPubMed
  8. 8.↵
    Benjafield AV, Ayas NT, Eastwood PR, et al. Estimation of the global prevalence and burden of obstructive sleep apnoea: a literature-based analysis. Lancet Respir Med. 2019;7(8):687–698.
    OpenUrl
  9. 9.↵
    Punjabi NM. The Epidemiology of Adult Obstructive Sleep Apnea. Proceedings of the American Thoracic Society. 2008;5(2):136–143.
    OpenUrlCrossRefPubMed
  10. 10.↵
    Mirer AG, Young T, Palta M, Benca RM, Rasmuson A, Peppard PE. Sleep-disordered breathing and the menopausal transition among participants in the Sleep in Midlife Women Study. Menopause. 2017;24(2):157–162.
    OpenUrlCrossRefPubMed
  11. 11.↵
    Jehan S, Zizi F, Pandi-Perumal SR, et al. Obstructive Sleep Apnea and Obesity: Implications for Public Health. Sleep Med Disord. 2017;1(4).
  12. 12.↵
    Drager LF, Togeiro SM, Polotsky VY, Lorenzi-Filho G. Obstructive Sleep Apnea. Journal of the American College of Cardiology. 2013;62(7):569–576.
    OpenUrlFREE Full Text
  13. 13.↵
    Monahan K, Redline S. Role of obstructive sleep apnea in cardiovascular disease. Current Opinion in Cardiology. 2011;26(6):541–547.
    OpenUrlCrossRefPubMed
  14. 14.↵
    Mazzotti DR, Keenan BT, Lim DC, Gottlieb DJ, Kim J, Pack AI. Symptom Subtypes of Obstructive Sleep Apnea Predict Incidence of Cardiovascular Outcomes. Am J Respir Crit Care Med. 2019;200(4):493–506.
    OpenUrlCrossRefPubMed
  15. 15.↵
    Labarca G, Dreyse J, Salas C, Letelier F, Jorquera J. A Validation Study of Four Different Cluster Analyses of OSA and the Incidence of Cardiovascular Mortality in a Hispanic Population. Chest. 2021;160(6):2266–2274.
    OpenUrlCrossRefPubMed
  16. 16.↵
    Azarbarzin A, Sands SA, Stone KL, et al. The hypoxic burden of sleep apnoea predicts cardiovascular disease-related mortality: the Osteoporotic Fractures in Men Study and the Sleep Heart Health Study. European Heart Journal. 2019;40(14):1149–1157.
    OpenUrlCrossRefPubMed
  17. 17.↵
    Azarbarzin A, Sands SA, Younes M, et al. The Sleep Apnea–Specific Pulse-Rate Response Predicts Cardiovascular Morbidity and Mortality. American Journal of Respiratory and Critical Care Medicine. 2021;203(12):1546–1555.
    OpenUrlCrossRefPubMed
  18. 18.↵
    Mazzotti DR, Drager LF. Opportunities for Cardiovascular Benefits in Treating Obstructive Sleep Apnea in the Secondary Prevention Scenario. Am J Respir Crit Care Med. 2020;202(12):1622–1624.
    OpenUrlCrossRef
  19. 19.↵
    Yeghiazarians Y, Jneid H, Tietjens JR, et al. Obstructive Sleep Apnea and Cardiovascular Disease: A Scientific Statement From the American Heart Association. Circulation. 2021;144(3):e56–e67.
    OpenUrlCrossRefPubMed
  20. 20.↵
    Marin JM, Carrizo SJ, Vicente E, Agusti AGN. Long-term cardiovascular outcomes in men with obstructive sleep apnoea-hypopnoea with or without treatment with continuous positive airway pressure: an observational study. Lancet (London, England). 2005;365(9464):1046–1053.
    OpenUrl
  21. 21.
    Gottlieb DJ, Yenokyan G, Newman AB, et al. Prospective study of obstructive sleep apnea and incident coronary heart disease and heart failure: the Sleep Heart Health Study. Circulation. 2010;122(4):352–360.
    OpenUrlAbstract/FREE Full Text
  22. 22.
    Shahar E, Whitney CW, Redline S, et al. Sleep-disordered breathing and cardiovascular disease: cross-sectional results of the Sleep Heart Health Study. American Journal of Respiratory and Critical Care Medicine. 2001;163(1):19–25.
    OpenUrlCrossRefPubMedWeb of Science
  23. 23.
    Somers VK, White DP, Amin R, et al. Sleep apnea and cardiovascular disease: an American Heart Association/american College Of Cardiology Foundation Scientific Statement from the American Heart Association Council for High Blood Pressure Research Professional Education Committee, Council on Clinical Cardiology, Stroke Council, and Council On Cardiovascular Nursing. In collaboration with the National Heart, Lung, and Blood Institute National Center on Sleep Disorders Research (National Institutes of Health). Circulation. 2008;118(10):1080–1111.
    OpenUrlFREE Full Text
  24. 24.
    Nieto FJ, Young TB, Lind BK, et al. Association of sleep-disordered breathing, sleep apnea, and hypertension in a large community-based study. Sleep Heart Health Study. JAMA. 2000;283(14):1829–1836.
    OpenUrlCrossRefPubMedWeb of Science
  25. 25.
    Peppard PE, Young T, Palta M, Skatrud J. Prospective study of the association between sleep-disordered breathing and hypertension. The New England Journal of Medicine. 2000;342(19):1378–1384.
    OpenUrlCrossRefPubMedWeb of Science
  26. 26.↵
    Li M, Hou W-S, Zhang X-W, Tang Z-Y. Obstructive sleep apnea and risk of stroke: a meta-analysis of prospective studies. International Journal of Cardiology. 2014;172(2):466–469.
    OpenUrlCrossRefPubMed
  27. 27.↵
    Yu J, Zhou Z, McEvoy RD, et al. Association of Positive Airway Pressure With Cardiovascular Events and Death in Adults With Sleep Apnea: A Systematic Review and Meta-analysis. JAMA. 2017;318(2):156–166.
    OpenUrlCrossRef
  28. 28.↵
    Javaheri S, Martinez-Garcia MA, Campos-Rodriguez F. CPAP Treatment and Cardiovascular Prevention. Chest. 2019;156(3):431–437.
    OpenUrlCrossRefPubMed
  29. 29.↵
    Pack AI, Magalang UJ, Singh B, Kuna ST, Keenan BT, Maislin G. Randomized clinical trials of cardiovascular disease in obstructive sleep apnea: understanding and overcoming bias. Sleep. 2021;44(2).
  30. 30.↵
    Wouters OJ, McKee M, Luyten J. Estimated Research and Development Investment Needed to Bring a New Medicine to Market, 2009-2018. Jama. 2020;323(9).
  31. 31.↵
    Fogel DB. Factors associated with clinical trials that fail and opportunities for improving the likelihood of success: A review. Contemporary Clinical Trials Communications. 2018;11:156–164.
    OpenUrlCrossRef
  32. 32.↵
    Robins JM, Hernán MA, Brumback B. Marginal structural models and causal inference in epidemiology. Epidemiology (Cambridge, Mass.). 2000;11(5):550–560.
    OpenUrl
  33. 33.↵
    Shi J, Norgeot B. Learning Causal Effects From Observational Data in Healthcare: A Review and Summary. Frontiers in Medicine. 2022;9.
  34. 34.↵
    Reynor A, McArdle N, Shenoy B, et al. Continuous positive airway pressure and adverse cardiovascular events in obstructive sleep apnea: Are participants of randomized trials representative of sleep clinic patients? Sleep. 2021.
  35. 35.↵
    Pépin J-L, Bailly S, Rinder P, et al. Relationship Between CPAP Termination and All-Cause Mortality. Chest. 2022;161(6):1657–1665.
    OpenUrlCrossRef
  36. 36.↵
    Wickwire EM, Tom SE, Vadlamani A, et al. Older adult US Medicare beneficiaries with untreated obstructive sleep apnea are heavier users of health care than matched control patients. Journal of Clinical Sleep Medicine. 2020;16(1):81–89.
    OpenUrl
  37. 37.↵
    Wickwire EM, Bailey MD, Somers VK, et al. CPAP adherence is associated with reduced risk for stroke among older adult Medicare beneficiaries with obstructive sleep apnea. Journal of Clinical Sleep Medicine. 2021;17(6):1249–1255.
    OpenUrl
  38. 38.↵
    Bock JM, Needham KA, Gregory DA, et al. Continuous Positive Airway Pressure Adherence and Treatment Cost in Patients With Obstructive Sleep Apnea and Cardiovascular Disease. Mayo Clinic Proceedings: Innovations, Quality & Outcomes. 2022;6(2):166–175.
    OpenUrlCrossRef
  39. 39.↵
    Waitman LR, Song X, Walpitage DL, et al. Enhancing PCORnet Clinical Research Network data completeness by integrating multistate insurance claims with electronic health records in a cloud environment aligned with CMS security and privacy requirements. J Am Med Inform Assoc. 2021.
  40. 40.↵
    Keenan BT, Kirchner HL, Veatch OJ, et al. Multisite validation of a simple electronic health record algorithm for identifying diagnosed obstructive sleep apnea. J Clin Sleep Med. 2020;16(2):175–183.
    OpenUrlCrossRef
  41. 41.↵
    Zhou Z, Rahme E, Abrahamowicz M, Pilote L. Survival Bias Associated with Time-to-Treatment Initiation in Drug Effectiveness Evaluation: A Comparison of Methods. American Journal of Epidemiology. 2005;162(10):1016–1023.
    OpenUrlCrossRefPubMedWeb of Science
  42. 42.↵
    Scharf MT, Keenan BT, Pack AI, Kuna ST. Mask Refills as a Measure of PAP Adherence. Journal of Clinical Sleep Medicine. 2017;13(11):1337–1344.
    OpenUrl
  43. 43.↵
    de Groot V, Beckerman H, Lankhorst GJ, Bouter LM. How to measure comorbidity. a critical review of available methods. J Clin Epidemiol. 2003;56(3):221–229.
    OpenUrlCrossRefPubMedWeb of Science
  44. 44.↵
    Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373–383.
    OpenUrlCrossRefPubMedWeb of Science
  45. 45.↵
    VanderWeele TJ, Ding P. Sensitivity Analysis in Observational Research: Introducing the E-Value. Annals of Internal Medicine. 2017;167(4).
  46. 46.↵
    Schwab RJ, Badr SM, Epstein LJ, et al. An official American Thoracic Society statement: continuous positive airway pressure adherence tracking systems. The optimal monitoring strategies and outcome measures in adults. Am J Respir Crit Care Med. 2013;188(5):613–620.
    OpenUrlCrossRefPubMedWeb of Science
  47. 47.↵
    Billings ME. Regional differences in PAP care: more questions than answers. Journal of Clinical Sleep Medicine. 2021;17(3):363–364.
    OpenUrl
  48. 48.↵
    Patel N, Sam A, Valentin A, Quan SF, Parthasarathy S. Refill Rates of Accessories for Positive Airway Pressure Therapy as a Surrogate Measure of Long-Term Adherence. Journal of Clinical Sleep Medicine. 2012;08(02):169–175.
    OpenUrl
  49. 49.↵
    Hu X, Fan J, Chen S, Yin Y, Zrenner B. The role of continuous positive airway pressure in blood pressure control for patients with obstructive sleep apnea and hypertension: a meta-analysis of randomized controlled trials. J Clin Hypertens (Greenwich). 2015;17(3):215–222.
    OpenUrlCrossRefPubMed
  50. 50.
    Bratton DJ, Gaisl T, Wons AM, Kohler M. CPAP vs Mandibular Advancement Devices and Blood Pressure in Patients With Obstructive Sleep Apnea: A Systematic Review and Meta-analysis. JAMA. 2015;314(21):2280–2293.
    OpenUrlCrossRefPubMed
  51. 51.↵
    Liu L, Cao Q, Guo Z, Dai Q. Continuous Positive Airway Pressure in Patients With Obstructive Sleep Apnea and Resistant Hypertension: A Meta-Analysis of Randomized Controlled Trials. J Clin Hypertens (Greenwich). 2016;18(2):153–158.
    OpenUrlCrossRef
  52. 52.↵
    Tung P, Levitzky YS, Wang R, et al. Obstructive and Central Sleep Apnea and the Risk of Incident Atrial Fibrillation in a Community Cohort of Men and Women. J Am Heart Assoc. 2017;6(7).
  53. 53.↵
    Marin JM, Carrizo SJ, Vicente E, Agusti AG. Long-term cardiovascular outcomes in men with obstructive sleep apnoea-hypopnoea with or without treatment with continuous positive airway pressure: an observational study. Lancet. 2005;365(9464):1046–1053.
    OpenUrlCrossRefPubMedWeb of Science
  54. 54.↵
    Javaheri S, Barbe F, Campos-Rodriguez F, et al. Sleep Apnea. Journal of the American College of Cardiology. 2017;69(7):841–858.
    OpenUrlFREE Full Text
  55. 55.↵
    Drager LF, McEvoy RD, Barbe F, Lorenzi-Filho G, Redline S, Initiative I. Sleep Apnea and Cardiovascular Disease: Lessons From Recent Trials and Need for Team Science. Circulation. 2017;136(19):1840–1850.
    OpenUrlAbstract/FREE Full Text
  56. 56.↵
    Campos-Rodriguez F, Queipo-Corona C, Carmona-Bernal C, et al. Continuous Positive Airway Pressure Improves Quality of Life in Women with Obstructive Sleep Apnea. A Randomized Controlled Trial. Am J Respir Crit Care Med. 2016;194(10):1286–1294.
    OpenUrlCrossRef
  57. 57.↵
    Engleman HM, Cheshire KE, Deary IJ, Douglas NJ. Daytime sleepiness, cognitive performance and mood after continuous positive airway pressure for the sleep apnoea/hypopnoea syndrome. Thorax. 1993;48(9):911–914.
    OpenUrlAbstract/FREE Full Text
  58. 58.↵
    Lyons OD, Ryan CM. Sleep Apnea and Stroke. Can J Cardiol. 2015;31(7):918–927.
    OpenUrlCrossRef
  59. 59.↵
    McEvoy RD, Antic NA, Heeley E, et al. CPAP for Prevention of Cardiovascular Events in Obstructive Sleep Apnea. N Engl J Med. 2016;375(10):919–931.
    OpenUrlCrossRefPubMed
  60. 60.
    Peker Y, Glantz H, Eulenburg C, Wegscheider K, Herlitz J, Thunstrom E. Effect of Positive Airway Pressure on Cardiovascular Outcomes in Coronary Artery Disease Patients with Nonsleepy Obstructive Sleep Apnea. The RICCADSA Randomized Controlled Trial. Am J Respir Crit Care Med. 2016;194(5):613–620.
    OpenUrlCrossRefPubMed
  61. 61.↵
    Sanchez-de-la-Torre M, Sanchez-de-la-Torre A, Bertran S, et al. Effect of obstructive sleep apnoea and its treatment with continuous positive airway pressure on the prevalence of cardiovascular events in patients with acute coronary syndrome (ISAACC study): a randomised controlled trial. Lancet Respir Med. 2020;8(4):359–367.
    OpenUrl
  62. 62.↵
    Sánchez-de-la-Torre M, Gracia-Lavedan E, Benitez ID, et al. Adherence to CPAP Treatment and the Risk of Recurrent Cardiovascular Events. Jama. 2023;330(13).
  63. 63.↵
    Reynor A, McArdle N, Shenoy B, et al. Continuous positive airway pressure and adverse cardiovascular events in obstructive sleep apnea: are participants of randomized trials representative of sleep clinic patients? Sleep. 2022;45(4).
  64. 64.↵
    Allen AJH, Jen R, Mazzotti DR, et al. Symptom subtypes and risk of incident cardiovascular and cerebrovascular disease in a clinic-based obstructive sleep apnea cohort. J Clin Sleep Med. 2022.
  65. 65.↵
    Wu B, Tarraf W, Wallace DM, et al. Cardiovascular correlates of sleep apnea phenotypes: Results from the Hispanic Community Health Study/Study of Latinos (HCHS/SOL). PLoS One. 2022;17(4):e0265151.
    OpenUrlCrossRefPubMed
  66. 66.↵
    Malhotra A, Sterling KL, Cistulli PA, et al. Dose-Response Relationship between Obstructive Sleep Apnea Therapy Adherence and Healthcare Utilization. Annals of the American Thoracic Society. 2023;20(6):891–897.
    OpenUrlCrossRef
  67. 67.↵
    Malhotra A, Cole KV, Malik AS, et al. Positive Airway Pressure Adherence and Health Care Resource Utilization in Patients With Obstructive Sleep Apnea and Heart Failure With Reduced Ejection Fraction. Journal of the American Heart Association. 2023;12(10).
  68. 68.↵
    U.S. Department of Health and Human Services Agency for Healthcare Research and Quality. Long-Term Health Outcomes in Obstructive Sleep Apnea: A Systematic Review of Comparative Studies Evaluating Positive Airway Pressure and the Validity of Breathing Measures as Surrogate Outcomes. 12/01/2022 2022. SLPT0919.
  69. 69.↵
    Ponce S, Pastor E, Orosa B, et al. The role of CPAP treatment in elderly patients with moderate obstructive sleep apnoea: a multicentre randomised controlled trial. European Respiratory Journal. 2019;54(2).
  70. 70.↵
    Martinez-Garcia MA, Oscullo G, Ponce S, et al. Effect of continuous positive airway pressure in very elderly with moderate-to-severe obstructive sleep apnea pooled results from two multicenter randomized controlled trials. Sleep Medicine. 2022;89:71–77.
    OpenUrlCrossRef
  71. 71.↵
    Greenberg-Dotan S, Reuveni H, Simon-Tuval T, Oksenberg A, Tarasiuk A. Gender differences in morbidity and health care utilization among adult obstructive sleep apnea patients. Sleep. 2007;30(9):1173–1180.
    OpenUrlCrossRefPubMedWeb of Science
  72. 72.↵
    Ayub S, Won CHJ. Obstructive Sleep Apnea in Women. Journal of Sleep Medicine. 2019;16(2):75–80.
    OpenUrlCrossRef
  73. 73.↵
    Simon-Tuval T, Reuveni H, Greenberg-Dotan S, Oksenberg A, Tal A, Tarasiuk A. Low Socioeconomic Status Is a Risk Factor for CPAP Acceptance Among Adult OSAS Patients Requiring Treatment. Sleep. 2009;32(4):545–552.
    OpenUrlCrossRefPubMedWeb of Science
  74. 74.↵
    Billings ME, Auckley D, Benca R, et al. Race and residential socioeconomics as predictors of CPAP adherence. Sleep. 2011;34(12):1653–1658.
    OpenUrlCrossRefPubMedWeb of Science
  75. 75.↵
    May AM, Patel SR, Yamauchi M, et al. Moving toward Equitable Care for Sleep Apnea in the United States: Positive Airway Pressure Adherence Thresholds: An Official American Thoracic Society Policy Statement. American Journal of Respiratory and Critical Care Medicine. 2023;207(3):244–254.
    OpenUrlCrossRef
  76. 76.↵
    Dalmar A, Singh M, Heis Z, et al. Risk of Atrial Fibrillation and Stroke After Bariatric Surgery in Patients With Morbid Obesity With or Without Obstructive Sleep Apnea. Stroke. 2021;52(7):2266–2274.
    OpenUrlCrossRefPubMed
  77. 77.
    Stewart NH, Walters RW, Mokhlesi B, Lauderdale DS, Arora VM. Sleep in hospitalized patients with chronic obstructive pulmonary disease: an observational study. Journal of Clinical Sleep Medicine. 2020;16(10):1693–1699.
    OpenUrl
  78. 78.↵
    Protasiewicz Timofticiuc DC, Vladu IM, Dtefan A-G, et al. Associations of Chronic Diabetes Complications and Cardiovascular Risk with the Risk of Obstructive Sleep Apnea in Patients with Type 2 Diabetes. Journal of Clinical Medicine. 2022;11(15).
  79. 79.↵
    Pépin J-L, Bailly S, Rinder P, et al. CPAP Therapy Termination Rates by OSA Phenotype: A French Nationwide Database Analysis. Journal of Clinical Medicine. 2021;10(5).
  80. 80.↵
    Pépin JL, Bailly S, Tamisier R. Big Data in sleep apnoea: Opportunities and challenges. Respirology. 2019:resp.13669.
  81. 81.↵
    Wickwire EM, Bailey MD, Somers VK, et al. CPAP adherence is associated with reduced inpatient utilization among older adult Medicare beneficiaries with pre-existing cardiovascular disease. Journal of Clinical Sleep Medicine. 2022;18(1):39–45.
    OpenUrl
  82. 82.↵
    Mokhlesi B, Punjabi NM. “REM-related” Obstructive Sleep Apnea: An Epiphenomenon or a Clinically Important Entity? Sleep. 2012;35(1):5–7.
    OpenUrlCrossRefPubMedWeb of Science
  83. 83.↵
    Varga AW, Mokhlesi B. REM obstructive sleep apnea: risk for adverse health outcomes and novel treatments. Sleep and Breathing. 2018;23(2):413–423.
    OpenUrlCrossRef
  84. 84.↵
    Kapur VK, Psaty BM. Obtaining Valid Estimates of the Effect of CPAP Therapy. Chest. 2022;161(6):1444–1445.
    OpenUrlCrossRef
Back to top
Posted February 15, 2024.
Download PDF

Supplementary Material

Email
Positive Airway Pressure Therapy Predicts Lower Mortality and Major Adverse Cardiovascular Events Incidence in Medicare Beneficiaries with Obstructive Sleep Apnea
Diego R Mazzotti, Lemuel R. Waitman, Jennifer Miller, Krishna M. Sundar, Nancy H. Stewart, David Gozal, Xing Song, Greater Plains Collaborative
medRxiv 2023.07.26.23293156; doi: https://doi.org/10.1101/2023.07.26.23293156
Twitter logo Facebook logo LinkedIn logo Mendeley logo
Citation Tools

Subject Area