Systemic confounders affecting serum measurements of omega-3 and -6 polyunsaturated fatty acids in patients with retinal disease

BACKGROUND Omega-3 polyunsaturated fatty acids (PUFAs) have a highly anti-angiogenic effect in animal models. However, the clinical relevance of omega-3PUFAs in human retinal pathologies remains unclear. The ARED 2 study found no effect of omega-3 PUFA supplementation on progression of age related macular degeneration (AMD). The aim of this study was to compare serum levels of omega-3- and omega-6 PUFAs between patients with diabetic retinopathy (DR), AMD and retinal vein occlusion (RVO), and to identify potential confounders of serum level measurements. METHODS Venous blood samples were collected from 44 patients with DR, 25 with AMD, 12 with RVO and 27 controls. The lipid phase was extracted and analyzed using mass spectrometry. Retinal disease staging was done by indirect funduscopy and FAG where appropriate. Patient demographics and medical history including current medication and fasting state were acquired. Tukey contrasts for multiple comparisons of the mean and linear regression analysis were used for statistical analysis. RESULTS Our data revealed no significant differences in omega-6 PUFA serum levels between patients with AMD, DR, RVO and controls (p > 0.858). Uncorrected omega-3 PUFA levels were significantly higher in patients with AMD compared to DR but not compared to controls (p = 0.004). However, after correcting for possible confounders such as body mass index (BMI), age, sex, fasting and use of statins, no statistically significant difference remained for serum omega-3 PUFA levels. Fasting was identified as an independent confounder of total omega-6 PUFAs, three individual omega-6 PUFAs and one omega-3 PUFA(p < 0.0427). Statin use was identified as an independent confounder of α-linolenic acid (an omega-3PUFA; p = 0.0210). CONCLUSION In this pilot study with relatively low patient numbers, we report significant differences in serum levels of omega-3PUFAs among patients with different types of retinal diseases. However, these differences were not robust for disease specificity after correction for possible confounders in our cohort. Our results demonstrate that serum lipid profiles need to be interpreted with caution since they are significantly altered by variables like fasting and medication use independent from the underlying disease. Correcting for respective confounders is thus necessary to compare serum lipid profiles in clinical studies.