The association of antidepressant medications and diabetic retinopathy among people with diabetes

Abstract

Objective

To determine if the use of antidepressants was associated with lower odds of diabetic retinopathy and if so, to determine if this association was mediated by decreased inflammation as measured by C-reactive protein (CRP).

Design

This was a cross sectional study of 1,041 participants with type 2 diabetes 40–85 years old from the 2005–2008 National Health and Nutrition Examination Survey (NHANES). Multiple logistic regression was used to examine the association between the outcome of diabetic retinopathy and the primary exposure of antidepressant medication usage. We also determined whether CRP meets standard criteria as a mediator between antidepressant use and diabetic retinopathy.

Results

Participants using antidepressants were less likely to have diabetic retinopathy (OR 0.50, 95% CI: 0.31–0.82). CRP did not meet one of the criteria for mediation. However, CRP was an effect modifier such that the association of antidepressant use and diabetic retinopathy was only present among participants with CRP ≥ 0.3 mg/dl. Among the antidepressant drug classes, selective serotonin reuptake inhibitor (SSRI) users had significantly lower odds of developing diabetic retinopathy compared to non-users of antidepressants.

Conclusions

Using representative survey data of US adults with type-2 diabetes, this study found that antidepressant use was associated with lower odds of diabetic retinopathy. Further longitudinal and experimental studies are necessary to confirm this finding and to determine if there is a role for antidepressants in preventing diabetic retinopathy in select patient populations.

Introduction

Diabetic retinopathy is the main cause of impaired vision among individuals between 25 and 74 years old in the United States (Zhang, Saaddine, Chou, et al., 2010). Although the pathogenesis of diabetic eye complications is not fully understood (Blum et al., 2012, Tomic et al., 2013), evidence suggests that inflammation and endothelial dysfunction may contribute to the development of retinopathy (Hamer et al., 2011, Lim et al., 2010, Nahomi et al., 1842). A hallmark of diabetic retinopathy is the formation of acellular capillaries that create local ischemia and lead to endothelial dysfunction in the retina. Pro-inflammatory signals are believed to drive histopathological changes in retina by promoting the synthesis of cytokines such as tumor necrosis factor α (TNF-α), interleukin-1β (IL-1β), and interferon-γ(INF-γ) (Tomic et al., 2013). Several studies, but not all, support the role of C-reactive protein (CRP) as an inflammatory marker associated with the development of retinopathy (Mima et al., 2012, Nguyen et al., 2009, Spijkerman et al., 2007, van Hecke et al., 2005, Walker, 2013).

Depression is common among patients with diabetes and associated with increased risk of diabetic complications (Sieu et al., 2011). Chronic, systemic inflammation may be an important contributor to the development of depressive symptoms (Lin et al., 2010, Miller et al., 2009, Raison and Miller, 2011). Several studies indicate that high systemic concentrations of CRP are associated with increased risk for psychological distress and depression (Douglas et al., 2004, Liu, Al-Sayegh, et al., 2014, Raison and Miller, 2011, Wium-Andersen et al., 2014). For this reason, one of the potential pharmacodynamic pathways for antidepressant medications is control of the immune/inflammatory response, which may lead to improvement in depressive symptoms and could also reduce the risk of developing retinopathy (Alboni et al., 2013, Crnkovic et al., 2012, Hamer et al., 2011, Vogelzangs et al., 2012). Alternatively, an indirect mechanism for antidepressants to prevent diabetic microvascular complications is their effects on glucose metabolism and glycemic control by improving medication adherence and health behaviors (Nicolau, Rivera, Frances, Chacartegui, & Masmiquel, 2013).

Anti-inflammatory effects of antidepressants have been studied over several decades. A recent study in male rats determined that cytokine gene expression was inhibited in the rat hypothalamus after a 28 day treatment with either fluoxetine (5 mg/kg) or imipramine (15 mg/kg) (Wang, Wang, Chen, Wu, & Xie, 2013). A shift in the balance of inflammation toward an anti-inflammatory state in the hypothalamus may represent a common mechanism of action for SSRIs and TCAs in animal research (Liu et al., 2014, Shenoy et al., 2013, Tomaz et al., 2014).

The efficacy of antidepressant therapy has also been manifested in changes in the concentration of CRP (Crnkovic et al., 2012, Vogelzangs et al., 2012). To examine the possible effect of antidepressant medications on inflammatory markers, 2,415 participants, including those with current depressive disorder, remitted depressive disorder, and healthy controls, were recruited in a cross-sectional study (Vogelzangs et al., 2012). The study found inconsistent associations between inflammatory markers and these three groups. A trend of higher CRP among TCA users vs. healthy controls was detected only in women. However, the level of interleukin-6 was significantly lower among men using an SSRI compared to healthy men (Vogelzangs et al., 2012). In summary, although some studies have suggested that antidepressant therapy reduces the concentration of inflammatory markers, this finding has been inconsistent among studies, and specifically, has not been conclusively demonstrated for CRP (Hamer et al., 2011, Hickman et al., 2014, Miller et al., 2009, Vogelzangs et al., 2012, Wium-Andersen et al., 2013).

Given that hyperglycemia induces tissue damage by inflammatory dysregulation (Lim et al., 2010, Nahomi et al., 1842), and based on the absence of studies exploring the effect of antidepressant medication on diabetic retinopathy, we aimed to determine whether the prevalence of retinopathy is associated with antidepressant use among diabetic adults. Given some published evidence that antidepressants reduce the concentration of CRP, we also evaluated the role of CRP as a mediator in this study.