Background To evaluate the additive effect of topical or sub-tenon injection of interferon (IFN)-alpha 2b in the treatment of refractory diabetic macular edema. Methods In this prospective study patients with center-involved DME who were unresponsive to 3 monthly consecutive IVB injections were recruited. Patients were divided into three groups: group1, received IFN- alpha 2b topical drop at a dose of 1mIU/ml four times a day for 3 months. Group 2, received a single sub-tenon injection of 1mIU/ml IFN- alpha 2b at the enrollment. Group 3 received artificial tears four times a day for 3 months (control group). All groups received three consecutive monthly IVB injections and were evaluated monthly up to 1 month following the last IVB injection. Results In this study, 59 eyes of 35 patients with refractory DME were assessed. The final follow-up showed that although CMT decreased in all groups, only patients in Group 2 had statistically significant lower CMT compared to their baseline values (change in CMT: - 117 +/- 213 mu m; p-value = 0.025). Comparison of CMT changes between three groups showed no statistically significant difference, although it was higher in group 2 (change in CMT: - 117 +/- 213 mu m (Group2) vs. - 49 +/- 173 (Group 1) vs. - 36 +/- 86 (Group 3); p-value = 0.085). Considering eyes with baseline CMT > 400 mu m, sub-tenon injection of IFN alpha 2b led to a significant reduction of CMT at the first month and final follow-up visit (CMT change: - 166 +/- 210, - 145 +/- 231 mu m; p-value = 0.018 and 0.035, respectively). In this subgroup, eyes in Group 2 had lower CMT at the first month following treatment in comparison with the control group (CMT: 444 +/- 123 mu m vs. 544 +/- 96 mu m, p-value = 0.042). Alterations of CDVA were not statistically significant among groups, although patients in Group 1 had a significant improvement in vision at second and last follow up (CDVA change: - 0.23 +/- 0.39, - 0.20 +/- 0.43 logMAR; p-value = 0.030 and 0.010, respectively). Conclusions In short term, Sub-tenon injection of IFN might have an additive anatomical effect in eyes with refractory DME. Validation of this observation requires further prospective controlled studies.