Sixty eyes of 60 patients with macular edema involving the fovea were enrolled in this study. Diffuse macular edema was defined by central thickening on a biomicroscopy using a 90-diopter noncontact lens and by diffuse fluorescein leakage on fluorescein angiography. Central macular thickness (CMT) was required to be >250 µm on an optical coherent tomography (OCT). Patients with other pathologies of macula or optic disc such as glaucoma or ocular hypertension were excluded. Patients with extensive foveal ischemia with more than one disc diameter of capillary closure on fluorescein angiography were excluded. Patients who had undergone intraocular surgery or macular grid laser photocoagulation within 3 months prior to the injection were excluded.

One of the authors (K.H.) performed intravitreal injection and another (J.O.) performed posterior subtenon injection during the same period. The physicians who checked visual acuity, IOP and CMT had not been informed of the purpose of this study or the assignment schedule.

Intravitreal injection was done under sterile conditions in the operating room. Eyes receiving the intravitreal injection were anesthetized by topical instillation of 0.5% proparacaine hydrochloride. Under the operating microscope, 0.1 ml of triamcinolone acetonide (4 mg) was injected slowly via a 30-gauge needle through the pars plana, 3.0 mm posterior to the limbus in the pseudopakic eyes and 4.0 mm posterior to the limbus in phakic eyes. For the posterior subtenon injection, under topical anesthesia (0.5% proparacaine hydrochloride), a small incision (8 mm posterior to the limbus) was made through the superotemporal forniceal conjunctiva and tenon's capsule to bare the sclera using Westcott scissors. The curved portion of the pinpoint cannula was inserted and a volume of 1 ml containing 40 mg triamcinolone was slowly injected behind the eye through the incision site. After the triamcinolone application was completed, the cannula was withdrawn slowly, with gentle pressure maintained by a sterile swab along the path of the cannula. No drug reflux was observed.

The patients were evaluated on the basis of central retinal thickness from OCT, visual acuity and IOP. All the patients underwent these examinations before injection and at 1 week, 1 month and 3 months post-injection. During the OCT procedure, each eye underwent six radial scans centered on the fovea. The value which printed on the OCT automated mode determined the central retinal thickness. The best-corrected visual acuity on the decimal charts was examined at each visit and was converted to the logarithm of minimal angle of resolution (logMAR) scale for statistical analysis. The IOP was measured using Goldmann applanation tonometry and the number of glaucoma medications used was recorded.

The temporal changes in logMAR visual acuity, central retinal thickness and IOP were compared using the Wilcoxon signed rank test. The differences between groups in logMAR visual acuity, central retinal thickness, IOP and other continuous variables were compared using the Mann-Whitney U test.

The mean age of the patients (±SD) was 60.69±10.8 years, with a range of 46 to 70 years. The mean (±SD) duration of diabetes was 16.3±8.1 years (range, 4-30 years). Patient demographics are shown in Table 1. No statistically significant differences were found between groups regarding age, gender, the ratio of left to right eyes, the ratio of PDR to NPDR or DM duration.

The mean (±SD) visual acuity before triamcinolone acetonide injection and at 1 week, 1 month and 3 months thereafter is shown in Table 2 and Fig. 1. The mean baseline visual acuity was not significantly different between the 2 groups (p=0.338). In both groups, the visual acuity significantly improved throughout the study. In the intravitreal injection group, the mean visual acuity at 1 month (0.571±0.393; p=0.001) and 3 months (0.535±0.382; p=0.001) after the injection were significantly better than baseline measurements. In the posterior subtenon injection group, the mean visual acuity at 1 month (0.650±0.281; p=0.011) and 3 months (0.623±0.264; p=0.007) after the injection was also significantly better than the baseline measurements. Between the 2 groups, there were no significant differences in the mean visual acuity changes before injection or at 1 month and 3 months after injection.

The mean baseline CMT was not significantly different between the 2 groups (p=0.461). In both the intravitreal and posterior subtenon injection group, the CMT was significantly decreased through the study (Table 3, Fig. 2). At 3 months after injection, OCT demonstrated a reduction of the mean CMT (intravitreal group: 46.2%, subtenon group: 43.5%). Between the 2 groups, there was no significant difference in the mean CMT change before injection or at 1 month and 3 months after injection. Fig. 3 illustrates the changes in the OCT images of a representative patient in the intravitreal injection group and of a patient in the posterior subtenon injection group.

The mean baseline IOP did not show differences between the 2 groups (p=0.524). In the intravitreal injection group, IOP tended to rise after the injection although the change was not statistically significant (Table 4, Fig. 4). Ten eyes (33%) experienced an IOP elevation to 21 mmHg or higher during the follow-up period. In the posterior subtenon injection group, a temporal change of IOP was not found. Only 1 eye (3.3%) experienced an IOP of 21 mmHg or higher during the follow up period. At 3 months after injection, the IOP change of the intravitreal injected eyes was greater than that of the posterior subtenon capsule injected eyes (p=0.026).

This study demonstrates that intravitreal injection or posterior subtenon injection has a beneficial effect in reducing the diabetic macular edema. Two groups did not show any significant difference in visual acuity or mean CMT thickness improvement after injection.

In this study, the change of mean visual acuity (LogMAR) after intravitreal injection was 0.160 (21.8%) at 1 month and 0.196 (26.8%) at 3 months after the injection. The change of mean CMT was 171.6 µm (40.1%) at 1 month and 197.7 µm (46.2%) at 3 months after intravitreal injection. These results were similar to previous studies. Martidis et al.5 reported that CMT decreased by 55% and 57.5% at 1 and 3 months after intravitreal trimcinolone injection, respectively. Ciardella et al.8 reported decreases of 42% and 46.4%. Jonas et al.9 reported that visual acuity(LogMAR) improved by 0.15 (15.3%) and 0.19 (19.3%) at 1 and 3 months after intravitreal trimcinolone injection, respectively.

Posterior subtenon injection of triamcinolone has been used with proven effectiveness in conditions with a breakdown in the blood-retinal barrier, such as cystoid macular edema and intermediate uveitis. Recently, Ohguro et al.14 reported an observational case series indicating the effectiveness of posterior subtenon triamcinolone infusion in diffuse diabetic macular edema in eyes that had not responded to vitrectomy. Bakri et al.15 also reported that visual acuities remained stable or improved over a 12-month period after posterior subtenon triamcinolone injections for refractory diabetic macular edema. These recent results have a general connection with our study.

Freeman et al.16 have shown by ultrasound B-scan that the superotemporal placement technique results in more accurate placement of steroids near the macula. Geroski et al.17 reported the usefulness of the transscleral pathway in delivering the drug to the retina. Weijtens et al.18 reported that the intravitreal concentration of the steroid increased after its peribulbar injection. On the basis of these reports, it can be summarized that the injected triamcinolone is located on the subtenon macular area and its therapeutic concentration on the choroids or retina can be obtained through the transscleral pathway.

Cardillo et al.19 compared intravitreal injection with posterior subtenon injection of triamcinolone in diabetic macular edema. They concluded that the intravitreal injection was more favorable than the posterior subtenon injection for the anatomic and functional aspects of improvement. Bonini-Filho et al.20 also suggested that intravitreal injection may be more effective than posterior subtenon injection for the management of refractive diffuse diabetic macular edema. These two studies differ from our results in that our study show that both intravitreal and posterior subtenon injections may be equally tolerated with a short-term performance. The study of Cardillo et al was informative in that two approaches were performed to each eye of the same patient with bilateral symmetric diffuse macular edema. However one limitation of the study was the relatively small sample size (12 patients). In Bonini-Filho et al's study, the patients had refractive diabetic macular edema. Thus, their results cannot directly be compared to our results.

One advantages of posterior subtenon administration is a lower risk of complication. IOP elevation is the most common complication after intravitreal triamcinolone injection.10-12 Although not statistically significant, IOP after intravitreal injection tended to rise in our study. At 3 months after injection, the change of IOP in the intravitreal injection group was greater than that of the posterior subtenon injection group. Other complications such as endophthalmitis and retinal detachment were also reported following intravitreal injection in other studies.

In conclusion, the short-term efficacy of the intravitreal injection and of the posterior subtenon injection of triamcinolone in diffuse diabetic macular edema was similar. The posterior subtenon injection was less invasive and safer than the intravitreal injection. Therefore, posterior subtenon injection of triamcinolone acetonide may be a good option for the treatment of diffuse diabetic macular edema.