Abstract
Purpose: :
To evaluate the impact of hemoglobin A1c (HbA1c) level as a moderator on bevacizumab effectiveness in patients with diabetic macular edema (DME).
Methods: :
This retrospective chart review included 45 eyes of 28 patients with diabetic retinopathy. Patients were included independent of the duration of diabetes, size of edema, visual acuity, age, sex, metabolic control, or previous treatments beyond one year. The treatment group (28 eyes of 20 patients with DME) received an initial 0.05 mL injection of 1.25mg of bevacizumab followed by a second dose administered an average of 53 days later. Central retinal thickness (CRT) measured by Optical Coherence Tomography (OCT) was recorded at the time of initial and follow-up injections, and an average of 54 days after the second injection. The control group included 16 eyes of 8 patients who did not receive bevacizumab but had CRT recordings at similar time intervals to the treated group: initial visit, second visit (mean 55.5 days later) and third visit (mean 48.8 days after second visit). The outcome measure was change in CRT between the first and third visits.
Results: :
The mean CRT of the treated group decreased by 56.46 (SD 145.43) microns while the mean CRT of the control group increased by 52.69 (SD 93.15) microns. In the treated group, using multiple regression analysis, we observed a linear relationship between the magnitude of CRT change and HbA1c of +27.2 (SE 12.6) microns per unit HbA1c (p=0.04). In the control group, this relationship was not significantly different from zero; the regression produced a coefficient of +2.1(p=0.7) with a 90% confidence interval of -30 to 10.9 (i.e. we can reject the hypothesis that the coefficient is greater than 10.9 with a p=0.05). To test the significance of HbA1c as a moderator variable on CRT response to bevacizumab, a moderated multiple regression (MMR) was performed. ANOVA indicated that the MMR model was significant (p=.0003). Furthermore, the HbA1c level was found to moderate the effectiveness of bevacizumab in reducing CRT (p=0.054). Specifically, the model estimates that the CRT change is related to HbA1c level in the treatment group by +25.5 (SE 12.6) microns per unit HbA1c versus -10.2 (SE 12.5) microns per unit HbA1c in the control group.
Conclusions: :
We observed a correlation between bevacizumab response measured by CRT and blood sugar control in patients with DME. Our study suggests that patients with poorly controlled diabetes do not respond as well to bevacizumab as patients with lower HbA1c.
Keywords: vascular endothelial growth factor • diabetic retinopathy • edema