Association Between Apolipoprotein E ∊4 and Sleep-Disordered Breathing in Adults

Context Apolipoprotein E ∊4 (ApoE ∊4) is a well-known risk factor for Alzheimer disease and cardiovascular disease. Sleep-disordered breathing occurs in Alzheimer disease patients and increases risks for cardiovascular disease. Complex interactions among sleep, brain pathology, and cardiovascular disease may occur in ApoE ∊4 carriers.

Objective To study whether genetic variation at the level of ApoE is associated with sleep-disordered breathing or sleep abnormalities in the general population.

Design, Setting, and Participants Ongoing longitudinal cohort study of sleep disorders at a US university beginning in 1989, providing a population-based probability sample of 791 middle-aged adults (mean [SD] age, 49 [8] years; range, 32-68 years).

Main Outcome Measure Nocturnal polysomnography to evaluate apnea-hypopnea index.

Results The probability of moderate-to-severe sleep-disordered breathing (apnea-hypopnea index ≥15%) was significantly higher in participants with ∊4, independent of age, sex, body mass index, and ethnicity (12.0% vs 7.0%; P = .003). Mean (SEM) apnea-hypopnea index was also significantly higher in participants with ApoE ∊4 (6.5 [0.6] vs 4.8 [0.3]; P = .01). These effects increased with the number of ApoE ∊4 alleles carried.

Conclusions A significant portion of sleep-disordered breathing is associated with ApoE ∊4 in the general population.

Sleep-disordered breathing (SDB) is prevalent but largely undiagnosed in adults.¹ Persons with SDB are at increased risk for hypertension^2,3 and have increased cardiovascular disease (CVD) morbidity and mortality.⁴ Characteristics of SDB, including changes in sleep architecture and electroencephalogram (EEG) slowing, are also present in persons with Alzheimer disease (AD).^5,6 Apolipoprotein E (ApoE), a protein involved in lipid metabolism, has 3 major allelic variants: ∊2, ∊3, and ∊4. While a protective effect for AD may be conferred by the ApoE ∊2 allele, risks for CVD and AD are increased by the ApoE ∊4 allele.^7,8 In this study, we assessed the contribution of ApoE genetic variation to SDB, sleep architecture, and other sleep parameters.

Data were obtained from participants in an ongoing longitudinal cohort study of sleep disorders that began in 1989.³ The cohort was constructed with a 2-stage probability sampling procedure on a random sample of employed men and women to maximize variability in SDB.^1,9 Every 4 years, participants underwent overnight polysomnography, blood sampling, and other tests. Informed consent was obtained in writing, using forms approved by the University of Wisconsin institutional review committee.

Studies involving participants who had fewer than 5 hours of polysomnographically documented sleep or who used psychotropic drugs were excluded (n = 558 studies). A total of 1344 overnight studies in 791 participants (up to 3 studies per participant) were included.

Sleep architecture and episodes of apnea and hypopnea were determined with standard in-laboratory polysomnography that included electroencephalography (EEG), electro-oculography, electromyelography, oximetry to detect arterial oxyhemoglobin saturation, thermistry and nasal pressure to detect airflow, and respiratory inductance plethysmography to record rib cage and abdominal excursions of breathing. Each 30-second epoch of the polysomnographic records was visually inspected and scored by trained technicians for sleep stage, apnea (≥10 seconds with no breathing), and hypopnea (a discernible reduction in the amplitude of respiratory inductance plethysmography associated with a ≥4% reduction in oxyhemoglobin saturation). The average number of apneas and hypopneas per hour of sleep (apnea-hypopnea index [AHI])³ was used as the summary measure of SDB. An EEG-slowing index (ratio of slow [delta and theta] to fast [alpha and beta] frequencies in eyes-closed, awake C3/A2 EEG) was calculated using fast Fourier spectral analysis⁶ on a subset of sleep studies with an adequate duration of quiet awake measurement (n = 381). Apolipoprotein E genotype was determined using the polymerase chain reaction–restriction fragment length polymorphism method.¹⁰ Serum cholesterol, triglyceride, and glucose levels were also measured.

Statistical techniques for data with repeated measures were used for our data of up to 3 studies per participant on factors with intraparticipant variation, including sleep architecture, AHI, and biochemical markers (SAS 8.0 software, SAS Institute, Cary, NC). Statistical analyses included repeated measures analysis of covariance regression for continuous outcomes (SAS PROC MIXED) and logistic regression for binary outcomes using the generalized estimating equations approach for repeated measures (SAS PROC GENMOD). Regression analyses were adjusted for potential confounding variables and for correlated observations within participants for multiple sleep studies. A binary outcome (AHI ≥15) was used to indicate clinically significant sleep apnea.

Allele frequencies for ApoE ∊2, ∊3, and ∊4 were 0.07, 0.78, and 0.15, respectively. Low-density lipoprotein (LDL) and triglycerides were increased while high-density lipoprotein (HDL) was decreased in ∊4-positive vs ∊4-negative participants (Table 1). Mean (SEM) total cholesterol and LDL were both decreased in ∊2-positive vs ∊2-negative participants (total cholesterol: 194 [3] mg/dL vs 206 [1] mg/dL, P<.001; LDL: 117 [3] mg/dL vs 131 [1] mg/dL, P<.001) (to convert mg/dL to mmol/L, multiply values by 0.0259). None of the other parameters differed between ∊2-positive and ∊2-negative participants (data not shown). A significant association between ∊4 and SDB was found. The prevalence of elevated (≥15) AHI and mean AHI were both significantly increased in ∊4-positive participants independent of age, sex, BMI, and ethnicity (Table 1 and Table 2). This association was present in both sexes (data not shown) and more pronounced in the 14 ApoE ∊4 homozygous participants of the cohort (24 sleep studies) (Table 2). Sleep architecture, EEG slowing (untransformed and log transformed) did not differ with ApoE ∊4 status (Table 1).

Our results indicate that ApoE ∊4 is associated with sleep apnea. We also found that ApoE ∊2 is associated with lower levels of total and LDL cholesterol while ∊4 is associated with higher levels of LDL and triglycerides, as previously reported.⁷ Sleep-disturbed breathing has been reported to cluster in families^11,12 and ∊4 might be 1 of the multiple genetic factors involved in susceptibility to this syndrome. Of note, SDB prevalence increases with aging,^1,13 and our sample is that of middle-aged adults. We found a substantial effect of ∊4 on SDB: there was a 2-fold increase in the odds of SDB (Table 2) in ∊4-positive vs ∊4-negative participants. Considering the prevalence of the ∊4 polymorphism (15%), up to 8% of AHI (15) in the general population might be caused by the effects of ∊4.

Only 1 study has examined the effect of ∊4 on SDB.¹⁴ In that study, participants with sleep apnea were more likely to be ∊4 homozygotes than were controls, but the difference was not statistically significant. The controls were mostly middle-aged men who had not been studied for SDB. Up to 9% of middle-aged men in the general population have an AHI of 15 or more,¹ so it is likely that the control group contained men with SDB, which would underestimate the association of ∊4 and SDB. Our study is unique because every participant has undergone nocturnal polysomnography, with most having been studied several times.

Another report suggests significant interactions among sleep, AD, and the ∊4 genotype, with a higher risk of AD morbidity present in ∊4-positive participants who napped for more than 60 min/d.¹⁵ Sleep-disordered breathing causes daytime sleepiness and prolonged napping. Thus, increased SDB in ∊4-positive participants might be responsible, at least partially, for this interaction.

Our finding is a simple statistical association that does not indicate a causal relationship between ApoE ∊4 and sleep apnea. A complex syndrome, SDB involves the central control of breathing and peripheral predisposing factors, leading to anatomical narrowing and collapse of the upper airway during sleep. Thus, SDB frequently occurs after such brain injuries as head trauma¹⁶ and stroke,¹⁷ demonstrating the importance of central factors. On the other hand, ∊4 may increase the density of β-amyloid deposits and neurofibrillary tangles in nondemented individuals.^7,8 Increased pathology in sleep/respiratory centers might contribute to centrally mediated SDB in ∊4-positive participants. Additional studies are needed to extend these findings.

The increased SDB prevalence in ∊4-positive participants may have clinical consequences. The established cardiovascular impact of SDB and the deleterious effect of ∊4 on lipid metabolism may have synergistic effects. Sleep-disordered breathing and ∊4 could also interact centrally to impair cognition. Not only may ∊4 predispose to neurodegenerative changes, but also SDB induces sleepiness and may damage the brain irreversibly through long-term hypoxemia.^18,19 Further studies will be needed to confirm and extend these findings.