Classification of Diabetic Retinopathy Severity with Aqueous VEGF Levels: A Prospective, Observational Study.

OBJECTIVE

A biomarker to classify diabetic retinopathy (DR) severity and guide preventative treatment is a major unmet need. We analyzed the relationship of aqueous VEGF-A with DR severity to determine if VEGF-A is a candidate biomarker.

DESIGN

A prospective, controlled trial at a tertiary academic medical center.

SUBJECTS: One hundred seven eyes of 107 patients without diabetes and 328 eyes of 164 adult type II diabetic patients with varying levels of DR.

METHODS

Aqueous was sampled in 107 consecutive eyes without diabetes undergoing elective vitrectomy surgery. All diabetic eyes had aqueous sampling of both eyes, ETDRS visual acuity, spectral-domain OCT, and color fundus photograph testing. Blood glucose and hemoglobin A1c (HbA1c) were measured. Aqueous VEGF-A was measured using a microparticle bead-based multiplex assay.

MAIN OUTCOME MEASURES: Aqueous VEGF-A, DR severity, and diabetic macular edema (DME).

RESULTS

Median and interquartile range (IQR) aqueous VEGF-A levels for the 107 eyes without diabetes was 55.74 pg/ml (23.56-76.72). Median (IQR) VEGF-A levels increased with DR severity: 100.34 pg/ml (72.31-126.07) in the no DR group, 162.69 pg/ml (113.76-233.89) in the moderate nonproliferative DR (NPDR) group, and 295.48 pg/ml (193.80-409.99) in the proliferative DR (PDR) group (P < 0.001). The moderate NPDR group was further subdivided based on ETDRS severity into mild-moderate NPDR (N = 154 eyes) and moderate-severe NPDR (N = 82 eyes). Median (IQR) VEGF-A levels were 147.24 pg/ml (108.15-187.50) and 206.75 pg/ml (136.18-303.25) in the mild-moderate and moderate-severe NPDR groups, respectively. Using the no DR group as reference, odds ratios with 95% confidence intervals for VEGF-A were 2.72 (1.3-5.7) in the mild-moderate NPDR (P = 0.008), 6.79 (2.75-16.73) in the moderate-severe NPDR (P < 0.001), and 15.02 (4.86-46.44) in the PDR (P < 0.001) groups, respectively. After controlling for gender, age, and HbA1c, increasing VEGF-A was significantly correlated with increasing DR severity (P < 0.001), but not with the presence of DME (P = 0.39).

CONCLUSIONS

Aqueous VEGF-A levels tightly correlate with DR severity and support VEGF-A as a novel biomarker to classify DR and guide early intervention. Use of VEGF-A levels may guide preventative treatment, reduce the risk of progression and vision loss, and allow optimal allocation of health care resources.

FINANCIAL DISCLOSURES: Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.