Magnesium status modulating the effect of serum vitamin D levels on retinopathy: National Health and Nutrition Examination Survey 2005 to 2008

Aim Magnesium levels may influence the effect of vitamin D levels on the body. This study aimed to assess the combined effect of magnesium status as reflected by magnesium depletion score (MDS) and vitamin D status on the risk of retinopathy. Methods This cross-sectional study included participants aged 40 years and older with complete information on vitamin D, MDS, and retinopathy assessment from the 2005–2008 National Health and Nutrition Examination Survey (NHANES). Logistic regression analysis was utilized to analyze the relationship of MDS and vitamin D with retinopathy and expressed as odds ratio (OR) and 95% confidence interval (CI). Results Of these 4,953 participants included, 602 (9.53%) participants had retinopathy. Serum vitamin D levels ≤30 nmol/L (vs. >30 nmol/L) (OR = 1.38, 95%CI: 1.05–1.81) and MDS >2 points (vs. ≤2 points) (OR = 1.47, 95%CI: 1.01–2.16) were associated with higher odds of retinopathy. There was an interaction between MDS and vitamin D on the increased odds of retinopathy (OR = 2.29, 95%CI: 1.12–4.68, Pinteraction = 0.025). In different MDS groups, serum vitamin D levels ≤30 nmol/L increased the odds of retinopathy only in the MDS >2 group (OR = 2.90, 95%CI: 1.16–7.24), but not in the MDS ≤2 group (p = 0.293). Subgroups analyses demonstrated that the interaction between MDS and serum vitamin D on retinopathy was observed in males (OR = 6.88, 95%CI: 1.41–33.66, Pinteraction = 0.019), people with diabetes (OR = 3.43, 95%CI: 1.78–6.63, Pinteraction < 0.001), and people with body mass index (BMI) ≥25 kg/m2 (OR = 2.46, 95%CI: 1.11–5.44, Pinteraction = 0.028). Conclusion Magnesium plays a moderating role in the relationship between serum vitamin D and retinopathy. The protective effect of vitamin D against retinopathy was primarily present among those with inadequate magnesium levels.


Introduction
Retinopathy is one of the major diseases that cause visual impairment and blindness, among which diabetic retinopathy is the leading cause of blindness in middle-aged and older adults worldwide (1).There is also a 6.7 to 18% prevalence of retinopathy in the population without diabetes, which may be related to advanced age and hypertension (2).Identifying modifiable factors that affect the risk of developing retinopathy is beneficial for disease prevention and reducing the burden of disease.
The retina is susceptible to oxidative stress (3).Vitamin D has been reported to prevent oxidative stress and inflammation in human retinal cells (4).Vitamin D may play a protective role in the retina through antioxidant, anti-inflammatory, anti-angiogenic, and immunomodulatory mechanisms (4,5).In addition, diabetes is one of the major risk factors for retinopathy, and vitamin D may protect the retina by improving insulin sensitivity and decreasing insulin resistance (6).Magnesium is an essential nutrient that plays an important role in the regulation of blood pressure, glucose metabolism, vascular tone (7,8), and it is involved in the synthesis and metabolism of vitamin D (9).Several studies have found that the effects of vitamin D on the body may vary depending on magnesium levels (10,11).For example, the relationship between serum 25-hydroxyvitamin D [25(OH)D] and the risk of death may be altered by the level of magnesium intake, and this negative correlation was found mainly in populations with higher magnesium intake (10).However, the joint effect of serum magnesium levels and vitamin levels on retinopathy is unclear.Furthermore, blood magnesium accounts for approximately 1% of whole-body magnesium, and although serum magnesium measurements can be used for the medical diagnosis of clinically severe magnesium deficiency, they do not reliably represent whole-body magnesium status (12,13).Since magnesium reabsorption in the kidney plays a crucial role in maintaining magnesium homeostasis (14), the magnesium depletion score (MDS) has been proposed as a new marker of magnesium status (15).MDS has been reported to be associated with self-reported risk of diabetic retinopathy (16).
Thus, this study aimed to assess the combined effect of magnesium status as reflected by MDS and vitamin D status on the risk of retinopathy in the middle-aged and elderly population, and to provide certain references for the prevention and management of retinopathy.

Study design and participants
The National Health and Nutrition Examination Survey (NHANES) dataset from 2005 to 2008 was used for this cross-sectional study.NHANES is an ongoing cross-sectional survey of health and nutrition of the United States noninstitutionalized population conducted by the National Center for Health Statistics (NCHS) of the Centers for Disease Control and Prevention (CDC). 1 The NHANES survey utilizes a complex multi-stage probability sampling design with

Assessment of retinopathy
Non-mydriatic digital images of the retina were captured from participants aged ≥40 years using the Canon CR6-45NM ophthalmic digital imaging system and Canon EOS 10D digital camera (Canon USA Inc., One Canon Park, Melville, New York).Two digital images were taken of each eye of the participants in an almost completely dark room, with the first image centered on the macula and the second on the optic nerve.Digital images were evaluated by graders at the University of Wisconsin according to a modified Airlie House classification system (18).Retinopathy severity was graded according to the Early Treatment Diabetic Retinopathy Study (ETDRS) grading scale (18).Participants with levels ≥14 were considered to have retinopathy according to the eye with the worse retinopathy level.The detailed assessment process is described in the NHANES database (19).
Serum vitamin D levels were obtained directly from NHANES records based on laboratory test data.Severe vitamin D deficiency with a serum 25(OH)D concentration below <30 nmol/L greatly increases the risk of mortality and many other diseases (20).In this study, vitamin D levels were categorized as ≤30 nmol/L and > 30 nmol/L.

Covariates
Participants' data were collected including age, gender, race, education, marital status, family poverty-to-income ratio (PIR),

Statistical analysis
Descriptive statistical analysis was performed in participants with and without retinopathy.Continuous data were described as mean and standard error (S.E.), and independent samples t-test was utilized to compare differences between the two groups.Categorical data were presented as frequency and percentage, and chi-square test or rank-sum test was used to compare differences between the two groups.
Variables with more missing values (e.g., physical activity, vitamin D intake, dialysis) were categorized as unknown, and variables with fewer missing values (<10%) were interpolated for missing values by the random forest multiple interpolation method using the "miceforest" package of the Python software.Difference analysis before and after missing value interpolation was performed (Supplementary Table S1).Weighted univariable logistic regression analysis was used to screen for covariates related to retinopathy (Supplementary Table S2).Weighted univariable and multivariable logistic regression analyses were utilized to assess the relationship of MDS and vitamin D with retinopathy: crude model was a univariable analysis; model 1 was a multivariable analysis that adjusted for age, gender, race, education, and PIR; model 2 was a multivariable analysis that adjusted for age, gender, race, education, PIR, diabetes, hypertension, CVD, CKD, dialysis, BMI, time of venipuncture, and vitamin D intake.The results were expressed as odds ratio (OR) and 95% confidence interval (CI).
The moderating effect of MDS on the relationship between serum vitamin D and retinopathy was analyzed.Crude model* included variables MDS, serum vitamin D, and interaction term "MDS × serum vitamin D. " Model 3 adjusted for age, gender, race, education, and PIR based on crude model*.Model 4 adjusted for age, gender, race, education, PIR, diabetes, hypertension, CVD, CKD, dialysis, BMI, time of venipuncture, and vitamin D intake based on crude model*.The interaction term "MDS × serum vitamin D" was used to assess the moderating effect of MDS on the relationship between serum vitamin D and retinopathy.In addition, the effect of the association between serum vitamin D and retinopathy was stratified in two groups of MDS (MDS > 2 and MDS ≤ 2).Subgroups analyses were performed based on gender, age, diabetes, and BMI.Data cleaning and processing of missing values were performed using Python 3.9 (Python Software Foundation, Delaware, United States), and statistical analysis was performed using SAS 9.4 (SAS Institute Inc., Cary, NC, United States).All statistical tests were performed using two-sided tests, and a p-value of <0.05 was considered statistically significant.

Moderating effect of MDS on the relationship between serum vitamin D and retinopathy
Table 3 shows the effect of interaction term "MDS × vitamin D" on retinopathy.There was an interaction between MDS and vitamin D on the increased odds of retinopathy [crude model*: (OR = 2.33, 95%CI: 1.17-4.64),P interaction = 0.018; model 3: (OR = 2.76, 95%CI: 1.40-5.45),P interaction = 0.005; model 4: (OR = 2.29, 95%CI: 1.12-4.68),P interaction = 0.025].Figure 2 shows the interaction between MDS and serum vitamin D on retinopathy.The risk of retinopathy showed a relatively smooth trend with decreasing serum vitamin D levels in the MDS ≤2 group, whereas the risk of retinopathy showed a rapid increase with decreasing serum vitamin D levels in the MDS >2 group.These results suggest that MDS plays a moderating role in the relationship between serum vitamin D and retinopathy.Table 4 presents the relationship between serum vitamin D and retinopathy in different MDS groups.In the MDS ≤2 group, serum vitamin D levels ≤30 nmol/L (vs.>30 nmol/L) increased the odds of retinopathy only in univariable analysis (OR = 1.47, 95%CI: 1.05-2.07).In the MDS >2 group, serum vitamin D levels ≤30 nmol/L (vs.>30 nmol/L) was related to higher odds of retinopathy both in univariable analysis (OR = 3.43, 95%CI: 2.01-5.86)and multivariable analysis (OR = 2.90, 95%CI: 1.16-7.24).
Because of the effect of age on retinopathy, we excluded 167 participants aged 80 years and older and used data from the remaining 4,786 participants for sensitivity analysis.The results demonstrated that there was still an interaction between MDS and vitamin D on the increased odds of retinopathy [model 4: (OR = 2.48, 95%CI: 1.22-5.05),P interaction = 0.014] (Supplementary Table S3).Due to too much missing data for the variable dialysis and the importance of the effect of dialysis on magnesium levels, we performed a sensitivity analysis after excluding the variable dialysis (Supplementary Table S4).The results showed that there was still an interaction between MDS and vitamin D on the risk of retinopathy after the exclusion of the variable dialysis [model 4: (OR = 2.26, 95%CI: 1.11-4.58),P interaction = 0.025].

Moderating effect of MDS in different subgroups
The moderating effect of MDS on the association between serum vitamin D and retinopathy in different subgroups were presented in Table 5.In different subgroups, the interaction between MDS and serum vitamin D on retinopathy was observed in males (OR = 6.88, 95%CI: 1.41-33.66,P interaction = 0.019), people with diabetes (OR = 3.43, 95%CI: 1.78-6.63,P interaction < 0.001), and people with BMI ≥25 kg/m 2 (OR = 2.46, 95%CI: 1.11-5.44,P interaction = 0.028).In addition, there may be an interaction between MDS and serum vitamin D on retinopathy in people older than 60 years (OR = 2.13, 95%CI: 0.97-4.66,P interaction = 0.059).

Discussion
This study examined the relationship between magnesium status and serum vitamin D levels and retinopathy in people aged 40 years and older.Serum vitamin D levels ≤30 nmol/L and high MDS (magnesium deficiency) were associated with higher odds of retinopathy.Moreover, MDS plays a moderating role in the relationship between serum vitamin D and retinopathy, and the moderating effect of MDS was observed only in males, people with diabetes, people older than 60 years, and people with BMI ≥25 kg/m 2 .
Retinopathy is caused by microangiopathy involving small pre-capillary retinal arterioles, capillaries, and small veins (26).Injury is caused by microvascular leakage and microvascular occlusion resulting from rupture of the blood-retinal barrier (26).Several studies have reported the protective role of vitamin D in the development of retinopathy (4,5,27).Vitamin D may exert retinal protective effects through antioxidant, anti-inflammatory, antiangiogenic, and immunomodulatory mechanisms (5).Vitamin D deficiency has been found to be associated with many eye diseases, such as myopia, age-related macular degeneration, glaucoma, diabetic retinopathy, and dry eye (5).Vitamin D has antioxidant and antiinflammatory properties and plays a role in anti-angiogenesis, regulation of cell proliferation, differentiation, and apoptosis (28,29).In addition, vitamin D prevents oxidative stress and inflammation in human retinal cells and increases the cellular viability of retinal pigment epithelial cells and various tissues (4).This current study analyzed the relationship between serum vitamin D levels and magnesium status and retinopathy.Our results demonstrated that low vitamin D levels and high MDS were related to higher odds of retinopathy.In addition, MDS plays a moderating role in the effect of serum vitamin D on retinopathy.The risk of retinopathy changed insignificantly with decreasing serum vitamin D levels in the low magnesium depletion group, whereas the risk of retinopathy showed Magnesium plays an important role in maintaining normal metabolism and ionic balance in ocular tissues (30).Membraneassociated ATPases, enzymes for ATP production and hydrolysis are magnesium-dependent (31).In the presence of magnesium deficiency, insufficient activity of antioxidant enzymes leads to lipid peroxidation of polyunsaturated fatty acid-rich membranes by free radicals, thereby impairing retinal function (31).For diabetic retinopathy, insulin resistance decreases intestinal and renal tubular epithelial activity and reduces magnesium absorption by the intestinal and renal epithelium, resulting in low serum magnesium (32).Low serum magnesium levels can further exacerbate insulin resistance, and the two affect each other (33).Magnesium intake can reduce oxidative stress and improve insulin and glucose metabolism (34,35).In our further analyses, the moderating effect of MDS on the relationship between serum vitamin D and retinopathy was observed only in males, people with diabetes, people older than 60 years, and people with BMI ≥25 kg/m 2 .Sex differences in the moderating effect of MDS may be related to sex hormones.Serum magnesium concentrations have been reported to be positively correlated with estradiol (36).The moderating effect of MDS was significant in people with diabetes and people with BMI ≥25 kg/m 2 may be associated with insulin and glucose metabolism due to the role of  The interaction between MDS and serum vitamin D on retinopathy.MDS, magnesium depletion score.magnesium in insulin and glucose metabolism (34,35).Since magnesium status plays a moderating role in the effect of serum vitamin D on retinopathy, the corresponding mechanism of effect may need to be further explored.This study is the first to examine the interaction of magnesium and vitamin D status on the risk of retinopathy in the middle-aged and elderly population based on data from a large nationally representative sample.This study provides epidemiologic evidence for the effect of magnesium modulating vitamin D levels on retinopathy.However, some limitations of this study should be noted.First, this was a cross-sectional study that could not infer causality, and residual confounders may have biased the results.Second, the effects of MDS and vitamin D levels on different subtypes of retinopathy could not be assessed because of the lack of appropriate data.Third, some of the information, such as medical history and physical activity, was obtained through self-report, which may have information bias.

Conclusion
This study explored the joint effect of magnesium status and serum vitamin D levels on retinopathy in people aged 40 years and older.Magnesium levels may play a moderating role in the relationship between vitamin D and retinopathy.The protective effect of vitamin D against retinopathy was primarily present among those with inadequate magnesium levels.The mechanisms underlying the moderating effect of magnesium status on the relationship between vitamin D and retinopathy may need to be further explored.

FIGURE 1 Flow
FIGURE 1Flow chart of the study population.MDS, magnesium depletion score; eGFR, estimated glomerular filtration rate; NHANES, the National Health and Nutrition Examination Survey database.

FIGURE 2
FIGURE 2 1 https://www.cdc.gov/nchs/nhanes/index.htm a two-year survey cycle.NHANES collects data through interviews and physical examinations, including demographic, dietary, socioeconomic, and health-related data, as well as medical, physiologic measurements, and laboratory test data.This study was based on two NHANES survey cycles, 2005-2006 and 2007-2008, because only these two cycles included full information on retinopathy based on retinal imaging exam.Participants were included according to the following criteria: Only participants in NHANES aged 40 years and older were included in this study because two-field, non-mydriatic retinal photography was performed only on this age group.The NCHS Research Ethics Review Board approved all NHANES protocols and each participant provided written informed consent.

TABLE 1
Characteristics of participants with and without retinopathy.

TABLE 2
Association of MDS and vitamin D with retinopathy analyzed by logistic regression analysis.Crude model, univariable analysis.Model 1, multivariable analysis that adjusted for age, gender, race, education, and PIR; Model 2, multivariable analysis that adjusted for age, gender, race, education, PIR, diabetes, hypertension, CVD, CKD, dialysis, BMI, time of venipuncture, and vitamin D intake.

TABLE 3
Interaction between MDS and serum vitamin D on retinopathy analyzed by logistic regression analysis.Serum vitamin D (≤30, >30 nmol/L) and MDS (≤2, >2) were analyzed as categorical variables; MDS, magnesium depletion score; OR, odds ratio; CI, confidence interval; Crude model*, included variables MDS, serum vitamin D, and interaction term "MDS × serum vitamin D"; Model 3, adjusted for age, gender, race, education, and PIR based on crude model*; Model 4, adjusted for age, gender, race, education, PIR, diabetes, hypertension, CVD, CKD, dialysis, BMI, time of venipuncture, and vitamin D intake based on crude model*.

TABLE 5
Interaction between MDS and serum vitamin D on retinopathy in different subgroups analyzed by logistic regression analysis.

TABLE 4
The relationship between serum vitamin D and retinopathy in different MDS groups analyzed by logistic regression analysis., magnesium depletion score; OR, odds ratio; CI, confidence interval; Ref, reference; Crude model, univariable analysis.Model 1, multivariable analysis that adjusted for age, gender, race, education, and PIR; Model 2, multivariable analysis that adjusted for age, gender, race, education, PIR, diabetes, hypertension, CVD, CKD, dialysis, BMI, time of venipuncture, and vitamin D intake. MDS