The Sleep Apnea-Specific Hypoxic Burden Predicts Incident Heart Failure

Abstract

Background: Heart failure (HF) is a leading cause of morbidity and mortality and although it is linked to sleep apnea, which physiological stressors most strongly associate with incident disease is unclear. We tested whether sleep apnea-specific hypoxic burden (SASHB) predicts incident HF in two independent cohort studies.

Research question: In comparison with apnea-hypopnea index (AHI), how does sleep apnea-specific hypoxic burden predict incident HF?

Study design and methods: The samples were derived from two cohort studies: The Sleep Heart Health Study (SHHS), which included 4,881 middle-aged and older adults (54.4% women), age 63.6 ± 11.1 years; and the Outcomes of Sleep Disorders in Older Men (MrOS), which included 2,653 men, age 76.2 ± 5.4 years. We computed SASHB as the sleep apnea-specific area under the desaturation curve from pre-event baseline. We used Cox models for incident HF to estimate the adjusted hazard ratios (HRs) for natural log-transformed SASHB and AHI adjusting for multiple confounders.

Results: The SASHB predicted incident HF in men in both cohorts, whereas AHI did not. Men in SHHS and MrOS had adjusted HRs (per 1SD increase in SASHB) of 1.18 (95% CI, 1.02-1.37) and 1.22 (95% CI, 1.02-1.45), respectively. Associations with SASHB were observed in men with both low and high AHI levels. Associations were not significant in women.

Interpretation: In men, the hypoxic burden of sleep apnea was associated with incident HF after accounting for demographic factors, smoking, and co-morbidities. The findings Suggest that quantification of an easily measured index of sleep apnea-related hypoxias may be useful for identifying individuals at risk for heart disease, while also suggesting targets for intervention.

Keywords: apnea-hypopnea index; heart failure; polysomnography; sleep apnea; sleep apnea-specific hypoxic burden.

Figure 1

Flow diagram presenting ascertainment of the study samples. CVD = cardiovascular disease; MrOS = the Outcomes of Sleep Disorders in Older Men study; NSRR = the National Sleep Research Resource; PSG = polysomnogram; SHHS = the Sleep Heart Health Study.

Figure 2

Examples of sleep apnea-specific hypoxic burden (SASHB) for two individuals with different desaturation patterns and similar SASHBs. A and B show the overlaid oxygen saturation signals (SpO₂) associated with all respiratory events for these two individuals. These signals were ensemble-averaged to obtain subject-specific search window to calculate the individual areas under desaturation curves. SASHB was the sum of individual areas divided by total sleep time. As shown, different patterns of desaturation (eg, longer, deeper, and less frequent desaturations (A) vs shorter, milder, and more frequent desaturations (B) could result in similar SASHBs.

Figure 3

Unadjusted cumulative incidence Kaplan-Meier curves for quintiles of AHI (apnea-hypopnea index) and SASHB (sleep apnea-specific hypoxic burden) in men in the Sleep Heart Health Study (SHHS; A) and the Outcomes of Sleep Disorders in Older Men (MrOS; B) study.

Figure 4

Hazard ratio of incident heart failure (HF) in men in the Sleep Heart Health Study for 1-SD increase in the SASHB and AHI. Participants with prevalent HF were excluded. AHI and SASHB were natural log-transformed, scaled, and modeled separately. From 4,881 available participants in the analytical dataset, there were 17 missing BMI, 122 missing diabetes status, 55 missing COPD status, 14 missing smoking status, and 59 missing prevalent CHD status, resulting in 4,535 individuals in models 2a and 2b. This figure shows the effect size for men in SHHS when both men and women were included in a single model and sex interaction was added to the models. Models 3a and 3b show the effect sizes after excluding those with a central apnea index > 5 events/h. CHD = prevalent coronary heart disease including myocardial infarction, angina, and coronary revascularization. AHI = apnea-hypopnea index; HTN = hypertension; SASHB = sleep apnea-specific hypoxic burden; TST = total sleep time.

Figure 5

Hazard ratio of incident heart failure (HF) in men in the Outcomes of Sleep Disorders in Older Men study for the SASHB and AHI. Participants with prevalent HF were excluded. Coronary heart disease (CHD) included myocardial infarction, angina, and coronary revascularization. AHI and SASHB were natural log-transformed and modeled separately. From 2,653 available participants in the analytical dataset, there were 2 missing BMI, 14 missing smoking status, and 6 missing prevalent CHD status, resulting in 2,646 individuals in models 2a and 2b. Models 3a and 3b show the effect sizes after excluding those with a central apnea index > 5 events/h. See Figure 4 legend for expansion of abbreviations.

Figure 6

Hazard ratio of incident heart failure (HF) for categories of SASHB and AHI, compared with the baseline group of low SASHB and low AHI. This figure displays the pooled analysis of men in the Outcomes of Sleep Disorders in Older Men (MrOS) and the Sleep Heart Health Study (SHHS) for SASHB and AHI. Participants with prevalent HF were excluded. AHI (apnea and hypopneas associated with ≥4% desaturation) and SASHB were each categorized separately as low (<75th percentile) and high (>75th percentile), and the individuals with low AHI and low SASHB were considered the reference group. See Figure 4 legend for expansion of abbreviations.