Purpose To evaluate the effect of intravitreal triamcinolone acetonide (IVT) on refractory diabetic macular edema (DME). Methods In a prospective, placebo-controlled, randomized clinical trial, 88 eyes of 61 patients with clinically significant macular edema that would have responded unfavourably to laser photocoagulation were randomly assigned to two groups. The treatment group (45 eyes) received 4 mg IVT and the control group (43 eyes) received a placebo subconjunctival injection. The primary outcome was central macular thickness (CMT) measured by optical coherence tomography. Complete examination was repeated at 2 and 4 months post-intervention. Results The mean (SD) CMT before the intervention and at the 2- and 4-month follow-ups was 393 (151), 293 (109) and 362 (119) μm in the treatment group and 393 (166), 404 (134) and 405 (160) μm in the placebo group, respectively. The second month difference was significant (P = 0.01). The difference between visual acuity changes (0.15 logarithm of the minimum angle of resolution, logMAR) was significant at 2 months (P = 0.02) but reduced to 0.11 logMAR (P = 0.08) after 4 months. Reduction for hard exudates and petaloid pattern were significantly greater in cases at 4 months. Conclusions The therapeutic effect of IVT on DME is greatest at 2 months and decreases up to the fourth month post-intervention. However, in terms of cystoid macular edema and hard exudates, the effect is maintained up to 4 months.