Gender-specific association between circulating 25-hydroxyvitamin D levels and diabetic retinopathy in patients with type 2 diabetes mellitus

Abstract

Background: Diabetic retinopathy (DR), one of the most common microvascular complications of diabetes mellitus (DM), remains to be a major driver of vision loss worldwide. Vitamin D has been reported to be involved in DR pathogenesis, but results have been inconsistent. We aimed to explore the relationship between blood 25-hydroxyvitamin D, 25(OH)D, level and the risk of DR in patients with type 2 diabetes (T2DM).

Methods: A total of 535 adults with T2DM from our department were included. Demographic information, biochemical data, 25(OH)D, and sex hormones were collected. Fundus photography was performed to determine the presence of DR. Participants were grouped into the DR group and the non-DR (NDR) group according to the fundus examinations and four groups based on serum 25(OH)D levels as follows: normal, insufficient, deficient, and severely deficient. Multivariate logistic regression analysis was used to evaluate the association between 25(OH)D and risk of DR.

Results: Males but not females with DR had significantly decreased levels of 25(OH)D (16.4 ± 5.6 ng/ml vs. 21.0 ± 5.0 ng/ml, P = 0.001) and increased proportion of severe 25(OH)D deficiency (14.8% vs. 6.7%, P = 0.022) compared to those without DR. Likewise, there was a gradually increasing percentage of DR with the reduction of 25(OH)D levels only in males (35.7%, 44.4%, 53.2%, 70.3%, P = 0.022). Intriguingly, total testosterone (TT) levels decreased markedly in males with DR (12.9 ± 5.2 nmol/L vs. 14.2 ± 5.5 nmol/L, P = 0.035) compared to their counterparts and correlated positively with 25(OH)D (β = 0.161, P = 0.007), which did not occur in females. After multivariate adjustment, we observed a significant inverse association between serum 25(OH)D and the occurrence of DR in males, showing that the adjusted ORs (AORs) and 95% confidence interval for DR were 0.233 (0.070-0.779) in the normal group, 0.280 (0.103-0.756) in the insufficient group, and 0.477 (0.196-1.164) in the deficiency group with the severely deficient group as a reference (P_-trend = 0.003). However, such a significant association was not observed in females (P_-trend = 0.137).

Conclusion: We concluded a gender-specific relationship between serum 25(OH)D and the incidence of DR in T2DM, supported by a significant inverse association between serum 25(OH)D and DR only in males, which could be mediated by a marked reduction in TT levels.

Keywords: 25-hydroxyvitamin D; diabetic retinopathy; gender-specific difference; total testosterone; type 2 diabetes.

Figure 1

Flowchart of the study cohort. Among the 659 potential adult patients, 124 who did not meet the inclusion criteria were excluded and finally 535 were enrolled in the current analysis. NDR, patients without diabetic retinopathy; DR, patients with diabetic retinopathy.

Figure 2

A significant decreased serum 25(OH)D levels (A) and increased percentage of severe 25(OH) D deficiency (C) were observed in male patients with DR than those without DR, which was not observed in females (B). Similarly, there was a significantly increasing trend in the percentage of DR with the reduction of serum 25(OH)D levels in the total cohort (P = 0.014) and males (P = 0.022) but not in females (P = 0.499) (D).

Figure 3

Linear regression analysis of serum 25(OH)D level with sex hormones in females and males. The 25(OH)D levels positively correlated with TT in males (B) but not in females (A). No any significant correlation was observed between 25(OH)D and E2 in either gender (C, D). TT, total testosterone; E2, estradiol; 25(OH)D, 25-hydroxyvitamin (D) P < 0.05 were considered to be statistically significant.

Figure 4

Unadjusted odds ratios (ORs) and 95% confidence intervals (CI) for DR based on serum 25(OH)D levels in females and males. The ORs (95% CI) for DR did not change significantly across the different 25(OH)D groups in females (P-trend = 0.214) (A) but reduced gradually across the different 25(OH)D groups in males (P-trend = 0.001) (B). P < 0.05 were accepted as statistically significant.