Neurological Institute Outcomes

Symptoms and polysomnogram (PSG)-based characteristics were examined in relation to incident atrial fibrillation. Patients (N=43,433, age>18) who underwent PSG between 2004 and 2015 were included. Clusters were identified using latent class analysis of 23 symptoms, ESS score, and 24 PSG measures. Demographics, comorbidity, medication use, and PAP use was accounted for in the analyses via adjustment and sensitivity analyses. Patients were age 51.8 ± 4.5 years, 51.9% male, 74.9% white, 9.7% had a baseline AF, and 8.9% developed incident AF over a follow-up period of 7.6 ± 3.4 years. Five clusters were identified:

• Long Sleep + REM subtype (N=26,809, 621.7%) which had the longest total sleep time, shortest REM latency, and most REM sleep;
• Short Sleep + NREM subtype (N=6,126, 14.1%) which had the shortest sleep time, longest REM latency, and least REM sleep;
• Hypopneas subtype (N=2,661, 6.1%) which had the most hypopneas;
• Apneas + Arousals subtype (N=4,592, 110.6%) which had the highest AHI and arousal index; and
• Hypoxic + Sleepy subtype (N=3,245, 87.5%) which had the worst hypoxia and highest ESS score.

The Hypoxic + Sleepy subtype had a 48% greater risk of developing AF (HR=1.48, 95% CI=1.34-1.64), the Apneas + Arousal subtype had a 22% greater risk of developing AF (HR=1.22, 95% CI=1.11-1.35), and the Short Sleep + NREM subtype had 11% greater risk of developing AF (HR=1.11, 95% CI=1.01-1.22) than the Long Sleep + REM subtype, which was used as reference. The Hypopneas subtype did not differ from reference.

Of the five distinct clusters identified, the sleep profile which conferred the strongest AF risk had the highest degree of hypoxia and dozing propensity. Illustrating clinical risk with a constellation of signs and symptoms, instead of solely the AHI for example, is a holistic and valuable prognostic tool paying the way for precision medicine.

An assessment was conducted of the association of sleep apnea across the socioeconomic spectrum and association with major adverse cardiovascular events (MACE) and mortality. Area Deprivation Index (ADI), a biomarker of neighborhood socioeconomic disadvantage, was calculated by national rank with higher quartiles reflecting greater deprivation (ADI-Q1, 2, 3, 4). Composite of MACE included heart failure, stroke, atrial fibrillation, and myocardial infarction. Statistical models took into account demographics, medications, and comorbidities.

Of 72,443 patients, those living in ADI-Q4, compared to those living in ADI-Q1, were more likely to be younger (48.9 ± 13.9 vs. 52.3 ± 14.0 years), female (59.9% vs. 40.1%), and African American (49% vs. 4.9%, all comparisons P < 0.001). MACE or death in ADI-Q4 was 28% greater than ADI-Q1 (HR=1.28, 95% CI=1.20-1.38, P <0 .0001). Likewise, increased AHI severity and percent time SaO₂<90% (T90) above median were associated with 13% and 21% increased risk of MACE or death (HR=1.13, 95% CI:1.07-1.19, P < 0.0001 and HR=1.21, 95% CI:1.15-1.27, P < 0.0001, respectively). Interactions were observed between T90 and ADI with the risk of MACE (P = 0.0016) or death (P = 0.0051). Living in areas of increasing socioeconomic deprivation confers an increased risk for MACE or death. Sleep-related hypoxia seems to modify this association.