ORIGINAL RESEARCH article

Front. Aging Neurosci., 22 August 2023

Sec. Neuroinflammation and Neuropathy

Volume 15 - 2023 | https://doi.org/10.3389/fnagi.2023.1247413

Sleep duration and age-related macular degeneration: a cross-sectional and Mendelian randomization study

  • 1. Department of Ophthalmology, The First Hospital of Wuhan, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, China

  • 2. Department of Neurology, The Third People’s Hospital of Chengdu, Chengdu, Sichuan, China

  • 3. Department of Ophthalmology, Yan’an People’s Hospital, Yan’an, Shaanxi, China

Abstract

Purpose:

To investigate the association between sleep duration and age-related macular degeneration (AMD).

Design:

Cross-sectional study, bidirectional two-sample Mendelian randomization (MR). For cross-sectional analysis, we used survey data of 5,481 participants aged ≥40 years from the 2005 to 2008 National Health and Nutrition Examination Survey (NHANES). For MR analysis, we used sleep- and AMD-associated genome-wide association studies (GWAS) data involving large populations.

Methods:

The association between sleep duration and AMD was assessed using logistic regression models. For MR analysis, the primary approach for MR analysis was the inverse-variance weighted (IVW) method.

Results:

In cross-sectional analysis, after adjusting for multiple covariates, short sleep duration (SSD) was found to be associated with increased risk of early AMD [odds ratio (OR) = 1.364, P = 0.036). MR analysis supported the results of cross-sectional analysis: SSD increases the risk of early AMD (β = 0.102, IVW-P = 0.003).

Conclusion:

Our findings provide the evidence supporting the association between sleep deficiency and higher risk of AMD. Further studies are required to confirm our findings and elucidate the mechanisms underlying this association.

1. Introduction

Age-related macular degeneration (AMD) is a neurodegenerative disease (NDD) involving neuroretina and retinal pigment epithelium (RPE), thereby leading to visual impairment or even blindness (Wong et al., 2014). Studies have reported considerable health burden in patients with AMD, which mainly affects adults aged 40 years and older (; Zhu et al., 2019). The number of patients with AMD is continuously increasing and estimated to be about 288 million worldwide by 2040 (; Wong et al., 2014). Previous studies have suggested that oxidative stress, inflammation and RPE senescence may all play a critical role (; ), however, the pathogenesis of AMD is still not fully understood.

Sleep disorders have been considered as a hallmark of aging and elders have increasing difficulty in falling and staying asleep with advancing age (). Particularly, sleep problems are frequently observed in patients with NDDs including Alzheimer’s disease (AD) (), and those with cardiovascular diseases such as hypertension and coronary artery disease (; ). Notably, AMD shares some common risk factors with these NDDs and cardiovascular diseases.

The National Health and Nutrition Examination Survey (NHANES) is a nationally conducted survey of the non-institutionalized civilian population in the United States. Previous studies have explored the risk or protective factors of AMD based on the NHANES data (; Zhu et al., 2020). In this study, we explored the association between sleep duration (SD) and AMD using the survey data from NHANES.

Observational studies are vulnerable to reverse causation, residual confounding, and selective bias (), which may negatively influence the estimates of relationship between SD and AMD. A randomized clinical trial (RCT) allows reliable and robust causal inferences to be drawn, but it is costly, time-consuming, and sometimes impractical to conduct one. Mendelian randomization (MR) approaches have opened up opportunities to assess and determine clinically associated characters for multiple diseases (), which examines causal relationships between exposures and outcomes using genetic variants significantly associated with an exposure as instrumental variables. By using MR approaches, these biases can be nicely overcome ().

Though there is increasing evidence indicating the correlation between sleep disorder and AMD, the underlying mechanisms are not identified yet and the evidence from observational studies is very limited. In this study, we aimed to comprehensively investigate the potential role of sleep abnormalities in the pathogenesis of AMD. As a result, cross-sectional and MR analysis indicated that sleep deprivation increased the risk of AMD. In conclusion, the results indicated that sleep deficiency could increase the risk of AMD.

2. Materials and methods

2.1. Cross-sectional study design

2.1.1. Participants

National Health and Nutrition Examination Survey data collected from two study cycles (2005–2006 and 2007–2008) was used in this study. Firstly, we identified total 7,081 participants aged 40 years and older. Subsequently, we excluded 1,253 participants without necessary retinal photographs and 337 participants without information on classification of AMD severity. Furthermore, 10 participants without sleep data were excluded. After that, the final 5,481 participants were identified (Figure 1). In the NHANES surveys, participants were asked to provide written informed consent before being enrolled.

FIGURE 1

2.1.2. Assessment of retinal photography and AMD severity

Retinal photographs of the final sample (5,481 participants) were collected. The assessment of the retinal photographs has been completed and provided on the NHANES website (). The diagnosis and classification of AMD was conducted by at least two experienced experts following a strict procedure. The definition of early AMD: signs of drusen with a grid area of >500 μm circle and/or pigmentary abnormalities. Late/advanced AMD was defined as the presence of exudative or geographic atrophy signs. If retinal photographs of both eyes were available, the status of the eye with higher AMD severity would be used in analyses.

2.1.3. Definition of sleep status

Abnormal SD was defined in the following way:

Insufficient/short sleep duration (SSD) (<7 h per night), normal (7–8 h per night), or excessive/long sleep duration (LSD) (≥9 h per night) (; ).

2.1.4. Covariates

The regressions models were adjusted for covariates that have been associated with SD and AMD: age, gender, race/ethnicity, education level, marital status, income level, alcohol consumption, smoking status, body mass index (BMI), self-reported history of hypertension, hyperlipidemia, diabetes mellitus, osteoporosis, general health condition, glaucoma, cataract surgery, and cardiovascular disease (CVD). Race were divided as non-Hispanic White and the others. Education level was categorized as those who had completed less than high school, completed high school or graduate equivalency degree, and completed more than high school. Income level is defined based on the poverty income ratio (PIR), with values <1.5 representing lower income, ≥1.5 and <3 representing middle income and ≥3 representing higher income. Alcohol consumption was categorized as no alcohol use, moderate (women: 1 drink per day or men: 1–2 drinks per day), heavy (women: 2–3 drinks per day or men: 3–4 drinks per day) and binge (women: ≥4 drinks per day or men: ≥5 drinks per day), according to definitions from the National Institute on Alcohol Abuse and Alcoholism. Smoking status was defined as low (blood cotinine <0.015 ng/ml), high (blood cotinine 0.015–3 ng/ml), and high (blood cotinine >3 ng/ml). BMI is calculated by the weight divided by height squared (kg/m2) and defined as not overweight (BMI <25 kg/m2), overweight (BMI 25–30 kg/m2) and obesity (BMI ≥30 kg/m2). Self-reported history of CVD was defined as having a previous physician diagnosis of congestive heart failure, coronary heart disease, angina, heart attack, or stroke.

2.2. Bidirectional MR study design

2.2.1. Data source for exposures and outcomes

In our study, we implemented bidirectional two-sample MR to judge causation using summary statistics from independent and large genome-wide association studies (GWAS) on SSD and LSD () and AMD (; Winkler et al., 2020). Summary-level statistics for early AMD from Winkler et al. (2020) contains 105,248 participants in European population (14,034 early AMD cases; 91,214 controls). The International Age-related Macular Degeneration Genomics Consortium (IAMDGC) contains 16,144 advanced AMD patients and 17,832 controls of European ancestry (). An ideal two-sample MR analysis requires a condition where no overlap between exposure and outcome participants occurs. Participants of exposures and outcomes are hardly overlapped and details regarding the characteristics of individual studies included in the consortium have been described in their website and the published paper (; ; Winkler et al., 2020).

2.2.2. Selection of genetic instruments

Instrumental variables (single nucleotide polymorphisms, SNPs) were selected via the following criteria: (i) with genome-wide significance (P < 5 × 10–8) and (ii) pruned by linkage disequilibrium (r2 < 0.001 and within 10,000 kb from the index variant). MR Pleiotropy RESidual Sum and Outlier (MR-PRESSO) (Verbanck et al., 2018) was used to remove the underlying outlier SNPs, thereby enhancing the robustness of the results. The F statistics of these SNPs were also taken into consideration. PhenoScanner () is an online platform with comprehensive information about genotype-phenotype association. We examined whether the obtained instrumental SNPs were associated with the outcomes and the potential confounders and subsequently remove the associated ones. The flowchart of MR analysis is presented in Figure 2.

FIGURE 2

2.2.3. MR approaches

Mendelian randomization’s validity depends on the crucial assumption of no pleiotropy (). Therefore, we used five MR approaches [random-effect inverse-variance weighted (IVW), MR Egger, weighted median, simple mode, and weighted mode] to address the heterogeneity and pleiotropy effect in estimating. The IVW method () was taken as the primary one, whereas the other methods were used to improve the robustness of IVW-derived results as what they are designed for , . We performed Cochran’s Q test to evaluate the heterogeneity and detect pleiotropy (). Besides, the MR-PRESSO (Verbanck et al., 2018) test was used to further assess the robustness of the causal effect between SD and AMD and SNPs identified as outliers by MR-PRESSO were removed. We also conducted leave-one-out analyses and MR-Egger intercept test to assess horizontal pleiotropy for significant estimates. In addition, we utilized a funnel plot to detect possible directional pleiotropy, similarly to how publication bias is assessed in a meta-analysis.

2.3. Statistical analyses

Statistical analyses were performed using R software (version 4.0.1) and EmpowerStats software. Continuous variables were expressed as means ± standard deviations (SD). Categorical variables were summarized as the counts or percentages (%).

For analysis of clinical data, EmpowerStats software and logistic regression model was used to calculate adjusted odd ratios (OR) and 95% confidence intervals (CIs). Three models were evaluated for outcomes: model 1 was unadjusted; model 2 was adjusted for age and gender; model 3 was further adjusted for the rest of covariates. Adjusted P < 0.05 was taken as significant.

For MR analysis, we performed all the analyses in R (version 4.0.1) using the TwoSampleMR () and MR-PRESSO (Verbanck et al., 2018) R packages. All statistical tests are two-sided. P < 0.05 was considered as nominally significant and P < 0.05/6 (8.33 × 10–3) (Bonferroni-corrected P) was considered as significant.

3. Results

3.1. Cross-sectional analysis

The characteristics of the study population are presented in Tables 1, 2. Among the participants, 77.2% were white people and 52.8% were females. Approximately 55.8, 65.1, and 35.5% of the people reported that they had completed more than high school, got married, and overweight, respectively. Almost 39.8% of participants reported an SSD status, whereas only 6.8% of participants reported an LSD status.

TABLE 1

CharacteristicsOverall (N = 5,481)Short SD (N = 2,181)Normal SD (N = 2,929)Long SD (N = 371)P-value
Age (%)<0.001
 40–6475.2 (73.9–76.5)79.3 (77.4–81.1)74.9 (73.2–76.6)53.2 (47–59.3)
 65–7919.7 (18.6–20.9)16.5 (14.9–18.3)20.2 (18.7–21.8)34.6 (29.1–40.6)
 ≥805.1 (4.6–5.6)4.2 (3.5–5)4.9 (4.3–5.6)12.2 (9.3–15.8)
Gender (%)<0.001
 Male47.2 (45.5–48.9)49.4 (46.7–52.1)47.1 (44.8–49.4)34.8 (29.2–40.8)
 Female52.8 (51.1–54.5)50.6 (47.9–53.3)52.9 (50.6–55.2)65.2 (59.2–70.8)
Race (%)<0.001
 Non-Hispanic White77.2 (76–78.3)69.6 (67.4–71.7)81.6 (80.1–83)81.9 (77.8–85.3)
 Others22.8 (21.7–24)30.4 (28.3–32.6)18.4 (17–19.9)18.1 (14.7–22.2)
Education (%)<0.001
 Less than high school17.9 (16.8–19)19.6 (17.9–21.5)15.9 (14.5–17.4)25.3 (20.6–30.6)
 High school or graduate equivalency degree26.3 (24.9–27.9)28.1 (25.7–30.6)25.2 (23.3–27.2)26 (20.8–32.1)
 More than high school55.8 (54.1–57.4)52.3 (49.6–54.9)58.9 (56.7–61.1)48.7 (42.4–55)
Marital status (%)<0.001
 Unmarried34.9 (33.4–36.5)40.3 (37.7–42.9)31.1 (29.1–33.1)38.3 (32.5–44.3)
 Married65.1 (63.5–66.6)59.7 (57.1–62.3)68.9 (66.9–70.9)61.7 (55.7–67.5)
Income level (%)<0.001
 Lower income18.7 (17.6–19.8)22.2 (20.3–24.2)15.9 (14.6–17.4)22.9 (18.5–28)
 Middle income24.1 (22.7–25.6)24.1 (21.8–26.4)22.9 (21–24.8)36.2 (30.2–42.6)
 Higher income57.2 (55.5–58.9)53.7 (51–56.5)61.2 (59–63.4)40.9 (34.5–47.7)
Smoking status (%)<0.001
 Low20.3 (19–21.8)18.6 (16.5–20.9)21.7 (19.8–23.7)18.3 (13.6–24.1)
 Moderate54.2 (52.5–55.9)51.9 (49.2–54.7)55.5 (53.2–57.8)55.1 (48.6–61.4)
 High25.5 (24–27)29.5 (27–32)22.8 (20.9–24.8)26.6 (21.3–32.8)
Alcohol use (%)0.032
 Moderate59.2 (57.1–61.3)56.5 (53–59.9)60.7 (57.9–63.4)61 (52.3–69)
 Heavy30.1 (28.2–32.2)31.5 (28.3–34.9)29 (26.5–31.7)32.7 (25.1–41.3)
 Binge10.7 (9.4–12)12 (9.9–14.4)10.3 (8.7–12)6.3 (3.2–12.3)
BMI (%)<0.001
 <25 kg/m2, not overweight27.7 (26.2–29.2)23.8 (21.6–26.2)29.7 (27.6–31.8)32.3 (26.6–38.5)
 25–30 kg/m2, overweight35.5 (33.9–37.2)36.5 (33.9–39.1)35.1 (33–37.3)33.4 (27.7–39.5)
 ≥30 kg/m2, obesity36.8 (35.2–38.4)39.7 (37.1–42.4)35.2 (33–37.4)34.3 (28.6–40.5)
Hypertension (%)<0.001
 No59 (57.3–60.6)57.1 (54.5–59.7)60.9 (58.7–63.1)52 (45.7–58.2)
 Yes41 (39.4–42.7)42.9 (40.3–45.5)39.1 (36.9–41.3)48 (41.8–54.3)
Hyperlipidemia (%)0.059
 No51.3 (49.4–53.1)49.9 (46.9–52.8)52.7 (50.2–55.1)46.9 (40.2–53.8)
 Yes48.7 (46.9–50.6)50.1 (47.2–53.1)47.3 (44.9–49.8)53.1 (46.2–59.8)
Diabetes (%)<0.001
 No89.5 (88.5–90.4)87.3 (85.6–88.9)91 (89.8–92.1)88.3 (84.4–91.2)
 Yes10.5 (9.6–11.5)12.7 (11.1–14.4)9 (7.9–10.2)11.7 (8.8–15.6)
Cataract surgery (%)<0.001
 No91 (90.2–91.8)92.1 (90.8–93.2)91.4 (90.3–92.4)80.9 (76–85)
 Yes9 (8.2–9.8)7.9 (6.8–9.2)8.6 (7.6–9.7)19.1 (15–24)
Glaucoma (%)<0.001
 No95.7 (95.1–96.3)95.5 (94.3–96.4)96.5 (95.7–97.1)89.9 (85.6–93.1)
 Yes4.3 (3.7–4.9)4.5 (3.6–5.7)3.5 (2.9–4.3)10.1 (6.9–14.4)
Health condition (%)<0.001
 Good or very good81.6 (80.3–82.7)76.6 (74.4–78.6)85.5 (83.9–86.9)75.8 (70.2–80.7)
 Fair15.3 (14.2–16.4)19.4 (17.5–21.5)12.1 (10.9–13.6)19.2 (14.8–24.5)
 Poor3.7 (2.7–3.7)4 (3.2–5)2.4 (1.8–3.1)5 (3–8.3)
Osteoporosis (%)0.166
 No92.4 (91.5–93.2)92 (90.5–93.2)92.9 (91.7–93.9)90.3 (86.4–93.3)
 Yes7.6 (6.8–8.5)8 (6.8–9.5)7.1 (6.1–8.3)9.7 (6.7–13.6)
CVD (%)0.052
 No82.1 (80.8–83.3)81.8 (79.8–83.7)82.8 (81–84.3)77.5 (72.4–81.9)
 Yes17.9 (16.7–19.2)18.2 (16.3–20.2)17.2 (15.7–19)22.5 (18.1–27.6)

Demographic characteristics, health-related behaviors, and comorbidities of participants with and without abnormal sleep duration (SD).

BMI, body mass index; CVD, cardiovascular disease.

TABLE 2

CharacteristicsOverall (N = 5,481)No AMD (N = 5,052)Early AMD (N = 374)Late AMD (N = 55)P-value
Age (%)<0.001
 40–6475.2 (73.9–76.5)78 (76.7–79.2)39.3 (32.9–46.1)7.4 (2.8–18.1)
 65–7919.7 (18.6–20.9)18.5 (17.3–19.7)39 (33.2–45.1)28.1 (16.6–43.4)
 ≥805.1 (4.6–5.6)3.5 (3.1–4)21.7 (17.7–26.4)64.5 (49.5–77.1)
Gender (%)0.055
 Male47.2 (45.5–48.9)47.2 (45.5–49)49.4 (43.1–55.8)30.1 (19.1–44.1)
 Female52.8 (51.1–54.5)52.8 (51–54.5)50.6 (44.2–56.9)69.9 (55.9–80.9)
Race (%)<0.001
 Non-Hispanic White77.2 (76–78.3)76.5 (75.3–77.7)85.5 (81.7–88.6)94.4 (87–97.7)
 Others22.8 (21.7–24)23.5 (22.3–24.7)14.5 (11.4–18.3)5.6 (2.3–13)
Education (%)0.195
 Less than high school17.9 (16.8–19)17.6 (16.4–18.7)22.4 (17.9–27.6)20.8 (12.1–33.5)
 High school or graduate equivalency degree26.3 (24.9–27.9)26.3 (24.8–27.9)26 (20.9–31.8)31.3 (19.6–46.1)
 More than high school55.8 (54.1–57.4)56.1 (54.4–57.8)51.6 (45.3–57.9)47.9 (34.1–62)
Marital status (%)<0.001
 Unmarried34.9 (33.4–36.5)34.4 (32.8–36.1)38.8 (33.1–44.9)65.5 (50.9–77.6)
 Married65.1 (63.5–66.6)65.6 (63.9–67.2)61.2 (55.1–66.9)34.5 (22.4–49.1)
Income level (%)<0.001
 Lower income18.7 (17.6–19.8)18.2 (17.1–19.4)23.9 (19.2–29.4)35.9 (23.7–50.2)
 Middle income24.1 (22.7–25.6)23.7 (22.2–25.2)30.9 (25.3–37.1)32 (19.7–47.4)
 Higher income57.2 (55.5–58.9)58.1 (56.4–59.9)45.2 (38.5–52)32.1 (19.6–47.9)
Smoking status (%)0.026
 Low20.3 (19–21.8)20 (18.6–21.5)25.9 (20.7–31.9)17.6 (9.2–31.1)
 Moderate54.2 (52.5–55.9)54.3 (52.5–56.1)50.3 (43.9–56.7)69.7 (55.3–81.1)
 High25.5 (24–27)25.7 (24.2–27.3)23.8 (18.4–30.1)12.7 (5.9–25.2)
Alcohol use (%)0.005
 Moderate59.2 (57.1–61.3)58.5 (56.3–60.7)70.5 (62.2–77.6)68.5 (43.9–85.8)
 Heavy30.1 (28.2–32.2)30.6 (28.5–32.7)24.2 (17.6–32.3)13.2 (3.2–41.5)
 Binge10.7 (9.4–12)10.9 (9.6–12.3)5.3 (2.7–10.2)18.3 (6.6–41.5)
BMI (%)0.029
 <25 kg/m2, not overweight27.7 (26.2–29.2)27.5 (25.9–29.1)28.7 (23.1–35)39 (26–53.8)
 25–30 kg/m2, overweight35.5 (33.9–37.2)35.3 (33.6–37)38.3 (32.4–44.6)45.1 (31.5–59.5)
 ≥30 kg/m2, obesity36.8 (35.2–38.4)37.2 (35.5–38.9)33 (27.3–39.2)15.9 (8.4–27.9)
Hypertension (%)<0.001
 No59 (57.3–60.6)59.7 (58–61.4)49.6 (43.3–55.9)40.3 (27.3–54.7)
 Yes41 (39.4–42.7)40.3 (38.6–42)50.4 (44.1–56.7)59.7 (45.3–72.7)
Hyperlipidemia (%)0.721
 No51.3 (49.4–53.1)51.4 (49.5–53.3)50.4 (43.6–57.1)45.3 (30.4–61)
 Yes48.7 (46.9–50.6)48.6 (46.7–50.5)49.6 (42.9–56.4)54.7 (39–69.6)
Diabetes (%)0.336
 No89.5 (88.5–90.4)89.6 (88.6–90.5)87.8 (83.4–91.2)84.5 (70.1–92.7)
 Yes10.5 (9.6–11.5)10.4 (9.5–11.4)12.2 (8.8–16.6)15.5 (7.3–29.9)
Cataract surgery (%)<0.001
 No91 (90.2–91.8)92.4 (91.6–93.1)76.4 (71.3–80.8)35.6 (23.3–50.1)
 Yes9 (8.2–9.8)7.6 (6.9–8.4)23.6 (19.2–28.7)64.4 (49.9–76.7)
Glaucoma (%)<0.001
 No95.7 (95.1–96.3)96.1 (95.5–96.7)90.2 (85.9–93.3)84.5 (71.2–92.3)
 Yes4.3 (3.7–4.9)3.9 (3.3–4.5)9.8 (6.7–14.1)15.5 (7.7–28.8)
Health condition (%)0.135
 Good or very good81.6 (80.3–82.7)81.8 (80.5—83)76.9 (71.6–81.5)86.9 (73.8–94)
 Fair15.3 (14.2–16.4)15 (13.9–16.2)19.9 (15.6–25)13.1 (6–26.2)
 Poor3.7 (2.7–3.7)3.2 (2.7–3.8)3.2 (1.9–5.4)–
Osteoporosis (%)<0.001
 No92.4 (91.5–93.2)92.8 (91.9–93.6)87.3 (82.8–90.7)79.5 (65.1–89)
 Yes7.6 (6.8–8.5)7.2 (6.4–8.1)12.7 (9.3–17.2)20.5 (11–34.9)
CVD (%)<0.001
 No82.1 (80.8–83.3)83.1 (81.8–84.3)70.2 (64.4–75.5)52 (37.9–65.7)
 Yes17.9 (16.7–19.2)16.9 (15.7–18.2)29.8 (24.5–35.6)48 (34.3–62.1)

Demographic characteristics, health-related behaviors, and comorbidities of participants with and without age-related macular degeneration (AMD).

BMI, body mass index; CVD, cardiovascular disease.

Three logistic regression models were used to explore the relationship between SD and AMD. When taking AMD as the outcome, a significantly positive association were observed between SSD and AMD (OR = 1.364, 95% CI: 1.058–1.941) in the multivariable logistic model (Table 3). When abnormal SD was considered as the outcome, there was a significant association between AMD and SSD in model 1 (OR = 1.944, 95% CI: 1.418–2.666); however, after adjusting for multiple covariates (model 3), no significant association between AMD and SSD was observed (Table 4).

TABLE 3

VariablesModel 1a OR (95% CI)P-valueModel 2b OR (95% CI)P-valueModel 3c OR (95% CI)P-value
Short sleep duration1.977 (1.34–2.917)<0.0011.388 (1.179–1.969)0.0071.364 (1.058–1.941)0.036
Long sleep duration1.011 (0.777–1.316)0.9341.17 (0.936–1.463)0.1691.13 (0.901–1.417)0.29

Logistic regression models of abnormal sleep duration for age-related macular degeneration (AMD).

aModel 1: unadjusted. bModel 2: adjusted for age and gender. cModel 3: further adjusted for race/ethnicity, education level, marital status, income level, alcohol consumption, smoking status, body mass index (BMI), self-reported history of hypertension, hyperlipidemia, diabetes mellitus, osteoporosis, general health condition, glaucoma, cataract surgery, and cardiovascular disease (CVD).

TABLE 4

OutcomesModel 1a OR (95% CI)P-valueModel 2b OR (95% CI)P-valueModel 3c OR (95% CI)P-value
Short sleep duration1.944 (1.418–2.666)<0.0011.343 (0.958–1.883)0.0871.334 (0.947–1.878)0.099
Long sleep duration0.738 (0.572–0.951)0.0190.93 (0.702–1.232)0.6140.963 (0.733–1.267)0.79

Logistic regression models of age-related macular degeneration (AMD) for abnormal sleep duration (SD).

aModel 1: unadjusted. bModel 2: adjusted for age and gender. cModel 3: further adjusted for race/ethnicity, education level, marital status, income level, alcohol consumption, smoking status, body mass index (BMI), self-reported history of hypertension, hyperlipidemia, diabetes mellitus, osteoporosis, general health condition, glaucoma, cataract surgery, and cardiovascular disease (CVD).

3.2. MR analysis of SD and AMD

To determine the causality and direction of the association between SD and AMD, we performed bidirectional MR analysis. For MR analysis, 8 instrumental SNPs were selected for genetically predicting early AMD (Supplementary Table 1), 23 SNPs for advanced AMD (Supplementary Table 2), 117 SNPs for SSD (Supplementary Table 3), and 27 SNPs for LSD (Supplementary Table 4). The F statistics for all of these genetic instruments exceeded the threshold of 10 (), suggesting that they were strong enough for being instruments. Results showed that SSD causally increases the risk of early AMD (β = 0.102, 95% CI: 0.008–0.195, IVW-derived P = 0.003) (Figure 3A and Table 5). In addition, advanced AMD causally increases the risk of SSD (β = 0.002, 95% CI: 0.001–0.003, IVW-derived P = 6.70E-05) (Figure 3B and Table 5). These two MR estimates all passed Bonferroni correction (IVW-derived P < 8.33 × 10–3, significant). A previous study has used MR approaches to evaluate the effect of SSD and LSD on advanced AMD () and found no significant effects. The leave-one-out plots (Supplementary Figure 1), the funnel plots (Supplementary Figure 2), the MR-Egger intercept test, Cochran’s Q test, and MR-PRESSO global test (Table 5) all suggested no apparent heterogeneity or pleiotropy in these results.

FIGURE 3

TABLE 5

ExposureOutcomeβ95% CIIVW-derived PP-heterogeneityP-interceptP-global
SSDEarly AMD0.1020.008–0.1950.0030.290.8760.81
LSDEarly AMD0.361−0.682 to 1.4030.4980.3250.7930.347
Early AMDSSD−0.003−0.009 to 0.0020.2040.4220.6620.571
Advanced AMDSSD0.0020.001–0.0036.70E-050.4830.1720.469
Early AMDLSD0.002−0.003 to 0.0040.9370.1210.9130.32
Advanced AMDLSD−5.80E-05−9.15E-04 to 8.00E-040.8950.3810.9580.422

Mendelian randomization estimates for associations between abnormal SD and AMD.

SSD, short sleep duration; LSD, long sleep duration; AMD, age-related macular degeneration; CI, confidence interval; IVW, inverse variance-weighted; β, beta-coefficient; P-heterogeneity, P-value for heterogeneity using Cochran’s Q test; P-intercept, P-value for MR-Egger intercept; P-global, P-value for MR-PRESSO global test; PRESSO, Pleiotropy RESidual Sum and Outlier.

4. Discussion

Patients with NDDs such as AD often experience sleep disorders (), while the association between SD and AMD has not been sufficiently reported. Clinical studies on this topic failed to include large populations and some of them showed apparently contradictory results. reported a significant association between short SD and neovascular AMD through a case-control study, while found that SD was not associated with neovascular AMD and longer sleep duration (LSD) was associated with geographic atrophy secondary to AMD. Hence, the relationship between SD and AMD has not been established and is needed to be determined and confirmed.

In this study, by using survey data from NHANES, cross-sectional analysis revealed that there was a significant association between SSD between AMD. However, observational studies are prone to reverse causation, residual confounding, and selective bias (), which may negatively influence the estimates of relationship between SD and AMD. More importantly, observational studies can only draw an association but not causality and the direction. A RCT allows reliable and robust causal inferences to be drawn, but it is costly, time-consuming, and sometimes impractical to conduct one. MR approaches have opened up opportunities to assess and determine clinically associated characters for multiple diseases (), which examines causal relationships between exposures and outcomes using genetic variants significantly associated with an exposure as instrumental variables. By using MR approaches, the biases and limitations mentioned above can be nicely overcome (). In this study, we used the two-sample MR approaches to further explore and determine the relationship between SD and AMD. The results of MR analysis supported the results of cross-sectional analysis: SSD increases the risk of early AMD and advanced AMD could increase the risk of SSD.

Sleep deficiency has been reported to be associated with various diseases and the pathophysiological changes caused by it are gradually becoming a major public health and safety issue with high economic and social costs (Yang et al., 2022). Sleep deficiency has been associated with overactivation of microglia and neuroinflammation (). Apoptosis, an active cell death process modulated by genes in multicellular organisms, regulates the development of the body and maintains the stability of the internal environment. In recent years, studies have found that expression of apoptosis indices (such as Bax) are significantly elevated in sleep deficiency animals (; ). Photoreceptors and RPE cells are highly energy-consuming cells in human retina, therefore, they are very sensitive to metabolic disorders. Notably, many studies have confirmed a causal relationship between sleep deficiency and energy metabolism disorders (; ). In addition, have found that sleep deficiency leads to changes in mitochondrial enzymes and decreases mitochondrial bioenergy efficiency in Drosophila. In the brain, sleep deficiency decreases mitochondrial membrane excitability and promotes the release of Cyt C, thus resulting in neuronal apoptosis (Xie et al., 2020). Studies have indicated that sleep deficiency may affect immune cells and the production of cytokines and complement, thereby causing immune system dysfunction and immune system-associated diseases (; ). Significant change in TNF-α level in rats has been observed after 20 and 30 days of chronic sleep deficiency (). In a study of sleep deficiency in 13 healthy young adults, found that five nights of sleep deficiency increases lymphocyte activation and pro-inflammatory cytokine production. In summary, potential mechanisms underlying the association between SSD and AMD include: induction of neuroinflammation, activation of cell apoptosis, energy metabolism disorder, and immune system dysfunction.

Sleep-awake cycle is controlled by the suprachiasmatic nucleus (SCN) of the midbrain. SCN receives photoperiodic information from intrinsic photosensitive retinal ganglionic cells (ipRGCs) in retina, through the retinal-hypothalamic pathway, and subsequently regulates the sleep-awake cycle of human body (). reported the impaired pupil light reflex (PLR) and function of ipRGCs in patients with early and neovascular AMD. Similarly, found the dysfunction of ipRGCs in patients with advanced AMD. These studies indicated that the increasing risk of SSD observed in patients with advanced AMD might be caused by ipRGCs dysfunction. Another potentially contributing factor to sleep deprivation in patients with AMD might be the significant impairment in brain function. The association between brain dysfunction and eye disorders has been widely reported. For instance, higher rates of depression have been reported in patients with AMD (). Moreover, depression has been associated with the alteration of sleep patterns (). Cognitive impairment or brain dysfunction in patients with AMD undermines the life quality even leads to disability of patients, which may play a role in abnormal SDs observed in our study.

This study also has some limitations. Although we tried our best to exclude and correct potential confounders, there still were some additional factors pertaining to sleep (e.g., whether participants were shift workers or had young children) that failed to be taken into consideration in cross-sectional analysis. In addition, light exposure is tightly associated with SD and might be the confounder when evaluate the SD-AMD association. However, this factor was not included for analysis due to loss of data. Furthermore, findings from this study only reveal a significant association between SSD and AMD, but not the underlying mechanisms, which call for further researches.

5. Conclusion

In conclusion, by cross-sectional and MR analyses, the results provide evidence that there is a significant association between SSD and higher risk of AMD. The findings could be a reminder that we ophthalmologists should pay more attention to individuals with sleep disorders. Further in-depth studies are required to confirm our findings and elucidate the mechanisms underlying this association.

Statements

Data availability statement

The original contributions presented in this study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.

Ethics statement

Human subjects or animal subjects were not included in this study. This study used only publicly available, deidentified summary statistics from previously published works, making it exempt according to the Yan’an People’s Hospital Institutional Review Board. Our research adhered to the tenets of the Declaration of Helsinki.

Author contributions

SL, HL, and ZL: conception and design. SL and ZL: data collection. SL: analysis and interpretation. HL: overall responsibility. All authors contributed to the article and approved the submitted version.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fnagi.2023.1247413/full#supplementary-material

Abbreviations

AMD, age-related macular degeneration; RPE, retinal pigment epithelium; MR, Mendelian randomization; SD, sleep duration; SSD, short sleep duration; LSD, long sleep duration; SNP, single nucleotide polymorphism; GWAS, genome-wide association study; NHANES, National Health and Nutrition Examination Survey; RCT, randomized clinical trial; PRESSO, Pleiotropy RESidual Sum and Outlier; IVW, inverse variance-weighted; OR, odds ratio; CI, confidence interval; IAMDGC, International Age-related Macular Degeneration Genomics Consortium; ipRGCs, intrinsic photosensitive retinal ganglionic cells.

References

Summary

Keywords

sleep, age-related macular degeneration, cross-sectional study, Mendelian randomization, causal effect

Citation

Lei S, Liu Z and Li H (2023) Sleep duration and age-related macular degeneration: a cross-sectional and Mendelian randomization study. Front. Aging Neurosci. 15:1247413. doi: 10.3389/fnagi.2023.1247413

Received

26 June 2023

Accepted

08 August 2023

Published

22 August 2023

Volume

15 - 2023

Edited by

Marios Kyriazis, National Gerontology Centre, Cyprus

Reviewed by

Gareth Banks, UK Dementia Research Institute, United Kingdom; Nava Zisapel, Tel Aviv University, Israel

Updates

Copyright

*Correspondence: Haihui Li,

Disclaimer

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.

Outline

Figures

Cite article

Copy to clipboard


Export citation file


Share article

Article metrics