In this cross-sectional study an increase in UA concentrations was independently associated with the presence of referable DR defined as a DR stage of severe non-proliferative or worse. For a clinical scenario, our study showed that values of UA of ≥ 7.8 mg/dL were independently associated with an increased risk of referable DR. Other factors that were independently associated were duration of DM, diastolic blood pressure and glomerular filtration rate.
Elevated UA is an important risk factor for progression of several lifestyle-related diseases such as metabolic syndrome and DM2 that often may have a common pathological basis (21). Hyperuricemia is frequently present in patients with metabolic syndrome. UA levels are determined by a balance between its production and excretion. The enzyme responsible for its production (xanthine dehydrogenase / xanthine oxidase) is expressed in the liver, the parenchyma cells of the small intestine, adipose tissue, macrophages and vascular endothelial cells (22). This enzyme is also a major source of oxidative stress and generation of reactive oxygen species.
It is proposed that microvascular damage of UA may be through inhibition of endothelial nitric oxide synthetase with a subsequent activation of the renin-angiotensin pathway (23). UA may also cause damage to endothelial cells and induce changes on vascular smooth muscle cells (24). Local production of UA in the vitreous cavity has been proposed as part of the pathogenic mechanisms of DR (25). UA has also been associated with increased vitreous levels of vascular endothelial growth factor in patients with diabetic macular edema (26). One possible explanation for the coexistence of both elevated UA and DR is that both reflect advanced stages of diseases that may often be present together in patients with poorly controlled metabolism. However, metabolism of UA may be linked to vascular endothelial cell damage and thus be directly related to the genesis and progression of DR.
Serum uric acid has been previously associated inconsistently with diabetic retinopathy. Xia et. al. found in a sample of 74 Chinese patients a higher concentration of UA between 39 subjects with DR and 34 without DR. In this study DR was classified as present or absent regardless of its degrees or severity (15). Lee et.al. found in a sample of 749 Taiwanese subjects followed for 3 years an increase in the severity of DR in 13.8% of the sample. Progression of DR was associated with variables like duration of DM, HbA1c, albuminuria and UA (14). Values of UA associated with progression analyzed with Cox regression were those of the 3rd (5.9–6.9 mg/dl) and 4th (≥ 7.0mg/dl) quartiles, with a hazard ratio (HR) of 2.57 (95% C. 1.30–5.08) and 3.66 (95% CI 1.92-7.00), respectively, when compared with the 1st quartile (< 4.9 mg/dl).
Kuwata et.al. studied in a sample of 1839 Japanese patients the association of baseline UA with the incidence rate of DR at 2 years of follow-up. A significant HR was found in men when the UA concentrations of the second (1.97 [95% CI, 1.14–3.41], p = .015), third (1.92 [95% CI 1.18–3.13], p = .008), and fourth quartiles, (2.17 [95% CI 1.40–3.37], p = .001) when compared with the first quartile. The association was not significant for women (16).
Previous studies have used different approaches in methods and statistical analysis for assessing the association between UA and DR. An important difference of our paper is that other studies used dichotomic definitions defined as present or absent for DR while in this paper we used referable retinopathy vs no referable retinopathy as the main outcome. This stage of DR has more clinical impact since it is related to potential visual loss and thus should be considered as more useful in the clinical practice (27).
In our study values of 7.8 mg/dL and above showed significant association values. These results suggest that higher levels of UA may be more suitable to be related to referable DR. Different to other studies we did not used quartiles as a mean for comparison. An advantage of using the threshold of 7.8 mg/dL is that it allows for the clinician a more practical approach to apply in the clinical setting with the patient. The normal values for UA have not been officially established, but findings in our study suggest a value close to 7.8 to be associated with the specific outcome of referable DR.
Avci et al. described a significant association of elevated UA with DR adjusted for DM duration (OR = 4.67, 95%CI 2.33–9.36). In our study the mean UA level was 5.4 mg/dl (IQR 4.5–6.6) in the group without DR and 6.5 mg/dl (IQR 5.8–8.1) in the DR group (p < 0.001). In Avci et al. study the main outcome was the presence of DR regardless of the severity, and no adjustment was made with other covariates (28).
Limitations of our study are inherent to a cross-sectional design. Also, being a hospital-based population limits the external validity of the findings. No adjustment by the use of UA lowering agents or medical therapy for hyperglycemia or hypertension was included in the analysis. This could lead to an under estimation of the association in the case that therapy could have lowered the high levels of glycemia, hypertension and UA. Despite the limitations, some strengths are worth mentioning, namely that other variables for DR were available to include in the logistic regression model, accounting for a more precise association.
In conclusion, our study suggest that UA may contribute to the microvascular damage in retinal vessels and therefore hyperuricemia could be a therapeutic target to prevent DR progression, although further longitudinal studies are needed to confirm or discard the importance of its treatment in the pathogenesis of DR.