Interactive correlations between artificial light at night, health risk behaviors, and cardiovascular health among patients with diabetes: A cross-sectional study

2.1 Survey participants

This is a cross-sectional study, which was based on the National Diabetes and Complications Surveillance Program. The analysis explores the relationship between ALAN and health risk behaviors (HRBs) and the prevalence of diabetes and its complications.

Our research methods and contents were conducted in accordance with relevant guidelines and regulations, and the informed consent of all participants was obtained. The study was carried out in accordance with the guidelines in the Declaration of Helsinki. All procedures involving participants were approved by the Ethics Committee of the First Affiliated Hospital of Anhui Medical University.

2.2 Sampling setting

In 2018, a nationwide screening for chronic complications of diabetes was implemented in 31 provinces/autonomous regions/municipalities, and the detailed sampling methods of the project were outlined before.²² Participants included people 18 years of age or older who had been diagnosed with diabetes and had lived in the survey area for more than 6 months in the 12 months prior to the survey. Pregnant women and individuals with mental illnesses were excluded.

2.3 Sample size estimation

2.4 Investigation contents and methods

Data were obtained via a questionnaire survey, body measurements, and laboratory testing. Sampling methods are described in detail in a previous study.²⁶

2.5 Outdoor light at night data

Outdoor ALAN data were obtained as per the protocol described in our previous study.²⁷ We utilized the highest quality global nighttime light map available, which is produced by the Earth Observation Group (https://eogdata.mines.edu/).^{28, 29}

2.6 CVH

CVH included central obesity, decreased high-density lipoprotein cholesterol (HDL-C), and elevated triglyceride (TG), blood pressure, and blood glucose, among others. Instances of two or more risk factors were designated “risk factor clusters.”³⁰

2.7 Statistical analysis

The SPSS 23.0 statistical software was used for analysis. Mean and standard deviation (SD) were used to describe continuous variables, and their association with outcomes was analyzed using univariate analysis of variance and T-test. In addition, frequency and percentage were used to describe categorical variables, and their association with outcomes were analyzed via chi-2 tests. Multivariate logistic regression was performed to analyze the correlation between HRBs and CVH, body mass index (BMI), waist circumference (WC), other metabolic indices (e.g., albumin [Alb], TG), and NDVI.

The relationship between ALAN and CVH was examined using several models: Model 1 evaluated the relationship between general demographic and CVH. Model 2 assessed the relationships between different HRBs, such as physical activity (moderate), sleep duration, metabolic markers, and CVH. Model 3 calculated complex metabolic measures, namely, clustered metabolic risk score (zMS), to summarize standardized values of WC, fasting TG, glycosylated hemoglobin (HbA1c), reciprocal systolic blood pressure, and HDL-C. These variables are standardized by subtracting the sample mean from the individual mean and then dividing by the standard deviation to get the corresponding value.³¹ The triglyceride glucose (TyG) index, which refers to a product of fasting blood glucose and triglycerides, was first introduced as a reliable alternative marker for insulin resistance.³² However, more recently, a close relationship has been established between insulin resistance and obesity, TyG-related parameters (including TyG), and anthropometric indicators of obesity (such as TyG-BMI and TyG-WC), which makes them superior to insulin resistance alone in predicting insulin resistance.³³ VAI: Male: [=(WC/[39.68 + 1.88 × BMI]) × (TG/1.03) × (1.31/HDL); Female: (WC/[36.58 + 1.89 × BMI]) × (TG/0.81) × (1.52/HDL)], where both TG and HDL levels were expressed in mmol/L.³⁴

The first step of our analysis involved descriptive statistics focusing on sociodemographic factors and CVH. Second, multilevel logistic regression was adopted to evaluate the association between ALAN, HRBs, and CVH. Third, an audit analysis was conducted using established procedures,³⁵ which was used to investigate the regulatory effect of ALAN, HRBs, and CVH. We also analyzed the relationship between ALAN and different types of HRBs, as well as that between ALAN, metabolic markers, and CVH. SPSS automatically calculates the interaction effect while generating the proportion of variance explained by the moderating effect of CVH on the relationship (evidenced by an increase in r²). Our model also accounted for sociodemographic correlations, and their effects were adjusted in the moderation analysis.

3.1 General demographic statistics

The general distribution of the variables examined among CVH is presented in Table 1. Among the 1765 participants for whom questionnaire responses were obtained, 906 (51.1%) were men, the average age was 59.10 ± 10.0 years, 853 (48.1%) lived in rural areas, 548 (30.9%) lived in towns, and 371 (20.9%) lived in cities. Among all assessed demographic details, sex, educational level, income level, and SES were significantly associated with CVH. By contrast, marital status, ethnicity, and age were not correlated with CVH.

3.2 Association between independent variables and CVH

The multilevel linear regression results and association with CVH are presented in Table 2. A dose–response relationship was observed between different ALAN and CVH (β = 0.205). After controlling for covariates, these relationships remained significant. A significant relationship between exercise (β = −1.557), vegetable intake (β = −1.022), walking (β = −0.449), and CVH was also established. Other results are reported in Table 2.

3.3 Moderation analysis between ALAN, exercise, and behaviors on CVH

Moderation analyses were performed for ALAN, exercise, behaviors, and CVH. The results are presented in Tables 3 and 4. We identified a positive association between exercise, vegetable intake, and CVH, with vegetable intake playing a moderating role in the increased CVH risk induced by ALAN (Table 3). Similar results were derived regarding ALAN, exercise, walking, and CVH (Table 4).

3.4 Moderation analysis between ALAN, exercise, and demographic information on CVH

Moderation analyses were performed for ALAN, exercise, demographic information, and CVH. Results suggest a positive association between ALAN exposure and CVH, with exercise and sex playing a moderating role in the increased risk of CVH induced by ALAN (Table 5). We observed similar results for ALAN, exercise, age, and CVH (Table 6), ALAN, exercise, SES, and CVH (Table 7), and ALAN, exercise, residential, and CVH (Table 8).

3.5 Moderation analysis between ALAN, exercise, and metabolic metrics on CVH

Moderation analyses were performed for ALAN, exercise, metabolic metrics, and CVH. Results suggest a positive association between ALAN exposure and CVH, with exercise and TG levels exhibiting moderating roles in the increased CVH risk induced by ALAN (Table 9). We obtained similar results for ALAN, exercise, TyG (Table 10), TG:HDL (Table 11), albumin (Alb; Table 12), VSI (Table 13), and CVH, as well as between ALAN, exercise, TyG-BMI index, and CVH (Table 14).

3.6 Moderation analysis between ALAN, NDVI, and metabolic metrics on CVH

Moderation analyses were performed for ALAN, NDVI, a metabolic metric (TG), and CVH. Among CVH, our results suggested a positive association between ALAN exposure and CVH; NDVI and TG play moderating roles in the increased risk of CVH induced by ALAN (Table 15). We observed similar results for ALAN, NDVI, vegetable intake, and CVH (Table 16).

3.7 Moderation analysis between ALAN, exercise, and CVH

Moderation analyses were performed for ALAN, drinking, walking, and CVH. The results are reported in Table S1. Among CVH, drinking and walking played moderating roles in the risk increase induced by ALAN. We observed similar results between ALAN, exercise, SES, and CVH (see Table S2).

3.8 Moderation analysis between ALAN, sleep duration, walking, and CVH

Moderation analyses were performed for ALAN, sleep duration, walking, and CVH (Table S3). Among CVH, sleep duration and walking played moderating roles in the risk increase of CVH induced by ALAN. We observed similar results in ALAN, sleep duration, sedentary behavior, and CVH (Table S4) as well as between ALAN, sleep duration, residential area, and CVH (Table S5).

3.9 Mediated moderation analysis between ALAN, sex, and CVH on CVD

Mediated moderation analyses were performed for ALAN, gender, CVH and CVD. The results are reported in Tables S6 and S7. Among CVD, gender plays a moderating role in the risk increase of CVD induced by ALAN mediated by CVH.

4.1 Principal results

This epidemiological survey examining diabetes among residents in four cities within Anhui Province established a positive correlation between exposure to residential outdoor ALAN and CVH. After adjusting for several significant diabetes-related risk factors, these associations remained robust. We identified certain modifiable factors that moderated the impact of ALAN on CVH among individuals with diabetes. Our results build upon previous research¹ and have important implications for evaluating the public health impact of light pollution on chronic diseases in China.

4.2 Comparison with previous research

Exposure to outdoor ALAN is a recognized risk factor for adverse health outcomes, including CVD.³⁶ A recent animal study suggests that short-term exposure to a common circadian disruptor, such as ALAN, increases inflammation and induces alterations in the angiogenic transcript within the mouse hippocampus.³⁷ Night shift work can significantly affect circadian rhythms owing to exposure to nighttime light, electronic devices, and other factors. Individuals working night shifts face a heightened risk of developing metabolic disorders and several types of cancer. This also reflects the risk of exposure to nighttime light in other population segments. Individuals exposed to ALAN, including during mealtimes, also exhibit disrupted circadian rhythms, increased metabolism, and heart disease. External factors that disrupt circadian rhythms can impair cellular functions and insulin sensitivity, leading to impaired glucose tolerance and lipid metabolism rhythms. Thus, disrupted circadian rhythms can have considerable detrimental effects on glucose and lipid metabolism, potentially increasing the risk of impaired glucose tolerance and the transition to diabetes and abnormal lipid metabolism.³⁸

Sleep is a vital biological process that can affect multiple functions, including cognition, development, energy conservation, and immune regulation.³⁹ Some research supports the strong association between ALAN and sleep as well as its harmful effects on sleep. ALAN is closely associated with sleep-related variables as it can disrupt the complete sleep cycle. The National China Child Health Study indicates that higher outdoor ALAN exposure levels are associated with a higher risk of sleep problems in children and are also more common in children under 12. This further suggests that effective control of outdoor ALAN exposure may be one of the key measures to improve children's sleep quality.⁴⁰ A cross-sectional study including older Chinese adults determined that an increase in nighttime light intensity located around the home outside was associated with the resulting decrease in sleep time.⁴¹ A cross-sectional and longitudinal study of more than 400 000 UK Biobank participants also clarified that exposure to daytime light, even at low levels, is an important environmental risk factor that affects mood, sleep, and circadian rhythm-related outcomes.⁴² A previous review highlighted the importance of causal relationships between outdoor ALAN, health, and behavior, and the importance of sleep as a key mediating factor,¹² thus suggesting that sleep can mediate complex relationships between variables. This also suggests that keeping bedrooms dark and reducing exposure to outdoor light may be feasible measures to reduce the prevalence of sleep problems.⁴³ A study comparing objective measures and self-reports documented that reported sleep deprivation was significantly associated with cardiometabolic risk factors for T2DM. This highlights the need for vigilance to reduce the adverse impacts of sleep deprivation and several other factors.⁴⁴

To further explore the role of HRBs in the development of CVH, a cross-sectional study of the relationship between light exposure, sleep patterns, physical activity, and changes in melatonin levels in 61 female shift nurses highlighted the importance of melatonin levels.⁴⁵ However, it did not establish a clear link between these variables, but determined that a consistent relationship between physical activity and melatonin, and sleep duration was not associated with urinary melatonin levels. These findings are particularly relevant because modifiable risk factors, such as changes in dietary and exercise patterns (low- to high-intensity exercise), are often the primary targets for addressing cardiometabolic problems.⁴⁶ A study examining the impact of shift work on CVD risk factors and MetS found that, even after accounting for potential covariates, such as job stress and dietary distribution, shift work was significantly associated with MetS.⁴⁷ This highlights the need to consider the metabolic effects of ALAN while also acknowledging the potential moderating roles of physical activity and sedentary behaviors.¹ Other studies suggest that physical activity is inversely associated with the incidence and progression of cardiometabolic diseases.^{48, 49} A prospective UK Biobank study revealed that long-term exposure to PM2.5 and NO₂ are linked to the prognosis of cardiometabolic multimorbidity. Older participants, men, individuals who drink excessive alcohol, and those with lower SES are more likely to experience these risks.⁵⁰ The US Center for Disease Control and Prevention reported a correlation between the Social Vulnerability Index and ALAN.⁵¹ To further clarify the interaction between ALAN and CVH in CVD, we conducted corresponding analyses, thus providing supplementary evidence to support the importance of considering the association between ALAN and CVD. Previous studies document that participants who live in areas with higher grass and tree cover have lower rates of MetS associated with ALAN,²¹ which is consistent with our results. This suggests the importance of developing green spaces, particularly in areas experiencing active or passive increases in ALAN.

The associated mechanisms include a direct effect of ALAN on the endocrine rhythms of the central nervous system, including the hypothalamic–pituitary–adrenal (HPA) axis and the hormone feedback mechanism that regulates its activity.⁵² From a possible mechanism viewpoint, ALAN is absorbed by various circulatory systems through the brain. Furthermore, hormones such as corticotropin releasing and arginine antidiuretic hormones are involved in the daily release of corticosterone and regulated by clock genes.⁵³ Melatonin is mainly involved in regulating the circadian cycle in the body, and its production and secretion are partly regulated by external light signals; therefore, it will also be affected by ALAN⁵⁴ and thus affect the secretion. Some studies indicate that the disruption of melatonin concentration is related to the changes of HPA axis rhythm,^{55, 56} thereby suggesting that ALAN exposure and the disruption of melatonin rhythm may further exacerbate neuroendocrine and metabolic disorders, resulting in a series of metabolic diseases. Fleury et al. reviewed animal experiments and population studies and revealed that exposure to ALAN causes circadian rhythm disruption, which negatively affects metabolic health.⁵⁷ Another perspective is the relationship between nighttime light exposure and altered metabolic function through changes in sleep. Since light exposure affects sleep, alterations in sleep patterns may be a primary mechanism accounting for the changes in glucose regulation.³⁶ ALAN disrupts the circadian rhythm to some extent by inhibiting the secretion of melatonin at night, and changes daily behavior such as exercise, sleep, and diet, which may lead to the development of metabolic diseases.⁵⁸ Detailed mechanisms between ALAN and circadian rhythms regarding cardiovascular and metabolic diseases including AMPK-OPA1-mitochondrial fusion/mitophagy axis, and circadian-related genes. Furthermore, the circadian clock plays a key role in linking obesity to the clock mechanisms of fat cells.⁵⁹ Another mechanism includes the association between ALAN and the inflammatory response, particularly in relation to the rhythm of inflammatory factors.^{60, 61} Animal experiments established that the expression of metabolic transcription factors Pparα and Pparγ increased in the epididymal fat of the low-dose ALAN group and the expression of Glut4 decreased in the heart.⁶² In addition to these harmful factors, the role of protective factors in CVH is also significant, including green spaces and healthy behaviors. These are viewed as potential mechanisms linking human health with the “active component” in nature. They have been recognized as having some degree of influence on health, including physical/mental states, behaviors, and conditions related to nature and health, including intense sleep and immune responses; specific health outcomes associated with nature (controlling for socioeconomic variables), including CVD,^{63, 64} are associated to some extent.

4.3 Limitations and strengths

The strengths of this study include results that align with those reported previously, specifically the interaction between ALAN and various behavioral patterns with CVH in patients with diabetes. Furthermore, we have developed corresponding countermeasures, thus providing a sound theoretical basis for further research focused on reducing ALAN exposure and effectively preventing the development of CVH through healthy behavioral interventions. Additional strengths include the multi-factor and multilevel study design, as well as the large number of participants, which facilitates comprehensive coverage of ALAN exposure levels across different cities.

Conversely, several limitations must be noted. First, the variables were all self-reported, and the imprecision and potential bias of their measurement associations limit our ability to further assess causality accurately. Second, our observational study faces challenges in establishing a causal relationship between ALAN, HRBs, and CVH. As an observational study, causation cannot be inferred. Thus, further cohort studies should be conducted to confirm and strengthen the potential association between outdoor ALAN exposure and disease risk across various groups. Third, we focused on such exposure without considering the effect of indoor lighting, particularly nighttime light exposure during sleep, on cardiometabolic outcomes. Finally, this study is limited to some cities in Anhui Province. Follow-up studies will be conducted with participants from different regions and cultures across the country. Since early detection and intervention are particularly effective in the treatment of future complications, this will include younger patients in particular. Future research will include additional mechanisms of circadian rhythm-related variables contributing to cardiac metabolic disease, prospective studies, and clinical trials.