Cardiology · Medical Literature
CPAP therapy may improve HFpEF-like features in coronary artery disease with obstructive sleep apnea: A randomized controlled trial.
Abstract
Obstructive sleep apnea (OSA) frequently coexists with coronary artery disease (CAD) and may contribute to heart failure with preserved ejection fraction (HFpEF). Whether HFpEF-like features are reversible with continuous positive airway pressure (CPAP) therapy remains uncertain. A total of 177 revascularized CAD patients with nonsleepy OSA (Epworth Sleepiness Scale score <10; apnea-hypopnea index [AHI] ≥15/h) from the RICCADSA cohort randomized to CPAP (n = 89) or no CPAP (n = 88) were included. HFpEF-like status was defined using a composite of five echocardiographic and biomarker criteria (≥2 positive). Multivariable logistic regression models were constructed to assess the association between CPAP therapy and HFpEF improvement, adjusting for age, sex, body mass index, AHI, hypertension, diabetes, log-transformed NT-proBNP, and E/e' ratio. HFpEF-like features were present in 126 patients (71.2%), with similar prevalence in the CPAP and no-CPAP groups (69.7% vs 72.7%). Among patients with HFpEF-like status at baseline, improvement at 12 month-follow-up occurred in 24 of 62 patients (38.7%) randomized to CPAP compared with 13 of 64 (20.3%) in the no-CPAP group (p = 0.023). In multivariable analysis, CPAP remained independently associated with HFpEF improvement (OR 2.85, 95% CI 1.13-7.22; p = 0.027). Higher baseline NT-proBNP was inversely associated with improvement (OR 0.37; p = 0.002), whereas E/e' was not independently predictive. CPAP therapy is independently associated with improvement in HFpEF-like features in patients with CAD and OSA, particularly among those with lower myocardial stress. These findings support the presence of a potentially modifiable HFpEF phenotype responsive to treatment of sleep-disordered breathing. ClinicalTrials.gov (NCT00519597).
Clinical relevance
Among revascularized patients with coronary artery disease and nonsleepy obstructive sleep apnea who have HFpEF-like echocardiographic and biomarker features at baseline, does 12 months of CPAP therapy increase the likelihood of improvement in those features compared with no CPAP? __SNAPSHOT__ {"studyDesign":"Randomized controlled trial (secondary analysis of the RICCADSA cohort); open-label with blinded outcome evaluation","setting":"Skaraborg County, West Sweden; two hospitals (Skövde and Lidköping) serving a regional population of approximately 250,000","sampleSize":"n = 177 total (CPAP: n = 89, no-CPAP: n = 88); HFpEF-like subgroup analyzed: CPAP n = 62, no-CPAP n = 64","population":"Revascularized CAD patients with nonsleepy OSA (Epworth Sleepiness Scale score <10; AHI ≥15 events/h) who also met the composite HFpEF-like definition (≥2 of 5 echocardiographic/biomarker criteria) at baseline","intervention":"Continuous positive airway pressure (CPAP) therapy for 12 months","comparator":"No CPAP (standard care)","primaryOutcome":"Improvement in HFpEF-like status at 12 months among patients with HFpEF-like features at baseline, defined as no longer meeting the composite HFpEF-like criteria (≥2 of 5 criteria: elevated filling pressures [E/e' ≥15], left atrial enlargement, increased LV mass index, elevated pulmonary artery systolic pressure [≥35 mmHg], and elevated NT-proBNP [≥125 pg/mL])","followUp":"12 months","fundingSource":"Not reported in the abstract; the parent RICCADSA trial has received Swedish governmental and research foundation support (not reported for this specific sub-analysis)"}
Key findings
Among the 126 patients (71.2% of the total cohort of 177) with HFpEF-like features at baseline, improvement at 12 months occurred in 24 of 62 patients (38.7%) randomized to CPAP versus 13 of 64 patients (20.3%) in the no-CPAP group. The absolute risk difference was 18.4 percentage points (calculated) in favor of CPAP. The relative risk of improvement was approximately 1.91 (calculated), suggesting CPAP-treated patients were roughly twice as likely to show improvement. The number needed to treat for one additional patient to show HFpEF-like improvement was approximately 6 (calculated; NNT = 1 / 0.184). A multivariable logistic regression adjusting for age, sex, BMI, AHI, hypertension, diabetes, log-transformed NT-proBNP, and E/e' ratio was performed; however, the adjusted odds ratio, 95% confidence intervals, and p-value for the primary outcome are not reported in the supplied abstract. Adverse events were not reported in the abstract.
Practice implications
This study provides randomized evidence suggesting that 12 months of CPAP may reverse composite HFpEF-like features in a meaningful proportion of revascularized CAD patients with nonsleepy OSA—approximately 38.7% vs. 20.3%, a difference that is both statistically noteworthy and potentially clinically important given the NNT of ~6 (calculated). However, several critical limitations must be acknowledged before translating these findings to practice. First, only the abstract was available for review; the adjusted effect estimate (odds ratio and 95% CI) from the multivariable logistic regression was not reported in the abstract, preventing full appraisal of the magnitude and precision of the treatment effect. Second, the HFpEF-like phenotype used here is a composite surrogate construct defined by echocardiographic and biomarker thresholds—it does not constitute a formal clinical HFpEF diagnosis, and improvement in the composite does not necessarily translate to reduction in clinical outcomes such as hospitalization, functional capacity, or mortality. Third, the parent RICCADSA trial demonstrated no significant reduction in hard cardiovascular events with CPAP in this same population, underscoring the gap between surrogate improvement and clinical outcome benefit. Fourth, participants were nonsleepy, a group known to have lower CPAP adherence (probability of remaining on CPAP at 2 years ~60% in the parent trial), which would attenuate any treatment effect in an intention-to-treat framework. Fifth, the cohort was predominantly male, post-revascularized, and recruited from a single Swedish county, limiting generalizability to women, non-European populations, or patients with non-obstructive CAD. Sixth, it is unclear whether the multivariable model was adequately powered given the sample size and event rate in the subgroup analysis. __LIMITATIONS__ Only abstract available; adjusted effect estimates (OR, 95% CI, p-value) for the primary logistic regression are not reported, precluding full statistical appraisal.,HFpEF-like status is a composite surrogate endpoint, not a formal clinical HFpEF diagnosis; improvement in the composite may not correspond to clinical benefit.,The parent RICCADSA trial failed to show reduction in hard cardiovascular events with CPAP in this same population, raising concerns about the clinical relevance of surrogate improvement.,Open-label design: patients and clinicians were unblinded to treatment allocation, introducing performance and ascertainment bias; echocardiographic readers may have been blinded, but this is not confirmed in the abstract.,Subgroup analysis (n=126 with HFpEF-like features; analytic n=62 vs. 64) is likely underpowered for the adjusted logistic regression, increasing risk of imprecise estimates.,Nonsleepy OSA patients have poor long-term CPAP adherence (~60% at 2 years in RICCADSA), which confounds interpretation of intention-to-treat results.,CPAP adherence data (hours/night) are not reported in the abstract, making it impossible to distinguish treatment effect by adherence level.,Predominantly male, post-revascularized Swedish cohort; limited generalizability to women, non-European populations, or medically managed CAD.,No information on adverse events or safety outcomes reported in the abstract.,Funding source for this sub-analysis is not reported, introducing potential for undisclosed conflicts of interest. __BOTTOM_LINE_CATEGORY__ promising_but_preliminary __BOTTOM_LINE__ In revascularized CAD patients with nonsleepy OSA and baseline HFpEF-like echocardiographic/biomarker features, 12 months of CPAP was associated with approximately twice the rate of composite HFpEF-like improvement compared with no CPAP (38.7% vs. 20.3%; NNT ~6, calculated), suggesting potential reversibility of these features. However, the adjusted effect estimate is not reported in the abstract, the endpoint is a surrogate composite rather than a clinical outcome, and the parent RICCADSA trial showed no hard cardiovascular event benefit with CPAP in the same population. These findings are hypothesis-generating and should not prompt changes in practice without replication in adequately powered trials using clinically meaningful endpoints.