Long-term Follow-up of Macular Translocation With 360° Retinotomy for Exudative Age-Related Macular Degeneration

Objective  To assess long-term functional and morphological changes after macular translocation in patients with exudative age-related macular degeneration.

Methods  Evaluation of a noncomparative cohort study of 90 patients with a follow-up of 14 to 79 months (mean, 38.2 months).

Results  Visual acuity increased by 3 or more lines in 15 patients, remained stable in 35 patients, and deteriorated in 40 patients at final examination. Pigment epithelium atrophy extending to the new fovea was detected in 44 patients; in 25 patients this new atrophy was associated with loss of visual function.

Conclusions  Long-term follow-up of macular translocation with 360° retinotomy showed stabilization or improvement in half of the patients. Progressive atrophy of the pigment epithelium represented one major limiting factor of the beneficial effect of the treatment. Macular translocation may be an option for cases of exudative age-related macular degeneration that are not eligible for or do not respond to alternative treatments.

Treatment of exudative age-related macular degeneration (AMD) is still controversial and the search for a standard procedure has not been successful. Laser photocoagulation and photodynamic therapy are effective therapeutic options for specific types.^1,2 Randomized studies testing the effect of intravitreal and systemic anti–vascular endothelial growth factor therapies are ongoing.^3,4 A number of studies have reported results after various surgical treatments of AMD, such as excision of choroidal neovascularization (CNV), transplantation of the pigment epithelium,^5-8 and full macular translocation.^9-15 The only large randomized multicenter trial of its kind, the Submacular Surgery Trial, showed no beneficial effect of CNV extraction on visual function compared with observation.¹⁶ The retinal pigment epithelium (RPE) plays a central role in maintaining visual function and its dysfunction may alter the extracellular environment of the photoreceptors, thus contributing to the pathogenesis and treatment failure of late-stage AMD. The purpose of full macular translocation is to move the fovea overlying the damaged or diseased RPE to a location with a healthier-appearing RPE. Several groups have reported recovery of distance and near visual acuity (VA) within the first year in an uncontrolled case series.^9-15 The purpose of this follow-up study was to evaluate long-term functional and morphological changes and to identify independent variables that might predict outcome after full macular translocation for neovascular AMD.

Between 1997 and 1999, 90 eyes of 90 patients with exudative AMD underwent a macular translocation operation in a prospective consecutive case series as published previously.¹³ All patients were contacted and scheduled for a long-term follow-up visit. General and ophthalmologic medical history was obtained from all patients since their last visit. Biomicroscopy of the anterior segment, ophthalmoscopy, intraocular pressure measurement, fundus photography, and fluorescein angiography were performed on all patients. Best-corrected distance VA was measured with the use of the Early Treatment Diabetic Retinopathy charts. Primary efficacy outcome was change in VA of the study eye at final examination compared with baseline. The secondary aim was to define factors influencing functional outcome. Data are described by means, SDs, and frequencies, as well as corresponding 95% confidence intervals. Univariate and multivariate analysis of covariance models were fitted to VA observation at final examination. Detailed comparisons were performed with 95% confidence intervals corresponding to the least significant difference t test. Effects with P < .05 were considered significant.

Mean follow-up was 38.2 months (range, 14-79). At the end of the observation period, 25 patients (27.8%) had died. Postoperatively, 63 patients (70%) were followed up for 2 years or longer, 52 patients (57.8%) for 3 years or longer, 38 patients (42.2%) for 4 years or longer, and 25 patients (27.8%) for 5 years or longer.

At 12 months after full macular translocation, mean VA had remained at 1.07 logarithm of the minimum angle of resolution (logMAR) units (20/200).¹³ Final VA ranged from 3.0 logMAR units (no light perception) to 0.3 logMAR units (20/40), with a mean of 1.26 logMAR units (20/300). At 1 year, an increase in VA of 3 or more lines was observed in 24 patients; VA was stable in 37 patients and deteriorated in 29 patients. At final examination, VA had increased in 15 patients, was stable in 35 patients, and had deteriorated in 40 patients (Table 1). Compared with a mean logMAR value of 1.00 (SD, 0.36) at baseline, mean logMAR value at 12 months was 1.07 (SD, 0.43) and 1.26 (SD, 0.53) at final examination, thus resulting in a mean change of 0.26 logMAR units (SD, 0.56) during the time period of about 3 years (Table 2). Figure 1 shows the survival curve with survival defined as not having lost 3 lines. At 12 and 36 months, 84% and 75% of patients, respectively, had not experienced a loss of 3 lines. Change in VA was influenced by the development of secondary RPE atrophy, including that of the new fovea. Mean change in VA in patients with central RPE atrophy was 0.37 logMAR units (SD, 0.60) compared with a change in VA of 0.0 logMAR units (SD, 0.36) in patients without RPE atrophy, a statistically significant difference (F_1.70 = 8.8; P = .004).

Univariate analysis showed that VA differences tended to be associated with type of CNV (F_2.87 = 4.88, P = .001), retinal complications (F_1.88 = 2.02, P = .16), and revisional operation (F _1.88 = 2.05, P = .16). No association with the surgeon (F_2.87 = 0.14, P = .87) or patient age (F_1.88 = 1.97, P = .16) was observed. Multivariate analysis of covariance gave a uniform assessment of the effects, indicating that VA differences were associated with baseline VA (F_1.83 = 9.79, P = .002) but not with length of follow-up (F _1.83 = 0.96, P = .33) (Figure 2), type of CNV (F_2.83 = 0.61, P = .54) (Figure 2), or revisional operation (F_1.83 = 2.35, P = .13).

A scatterplot of all patients (Figure 2) shows the distribution of change of patients' final VA as a function of the follow-up time. The distribution of the patients does not show better outcome at the shorter follow-up intervals or worse outcome at the longer intervals. Regression analysis of the change in VA as a function of CNV type did not show any linear trend.

Complications immediately following the operation and those occurring within the first year after the operation have been described previously.¹³ Between the 12-month and final examinations, 4 patients showed retinal complications with proliferative vitreoretinopathy retinal detachment that required additional surgical treatment. Late recurrence was seen in only 1 patient after the 1-year follow-up examination (at 51 months postoperatively); the patient was then treated by photocoagulation. Cystoid macular edema was diagnosed in 7 patients at final examination (in 3 eyes, cystoid macular edema persisted after 12 months; 4 eyes did not show cystoid macular edema at 12 months). At the final examination, excluding patients with detected early or late recurrence of CNV (Figure 3), secondary RPE and choroid atrophy extending to the new fovea was observed in 61% of patients (Figure 4). In 25 patients, foveal atrophy was associated with loss of 15 or more letters, indicating that secondary atrophy was 1 major factor in determining visual outcome after full macular translocation.

In this article, we present long-term results of a consecutive series of 90 patients who underwent full macular translocation for exudative AMD with a follow-up of about 3 years. At final examination, 54% of patients showed improved or stable VA and 46% had lost 3 or more lines. Our early enthusiasm about this therapy approach, which allowed for improvement of VA in some patients supported by encouraging results reported at 1 year after full macular translocation,^9-15 has been fading before the background of long-term results. Macular edema, proliferative vitreoretinopathy, and recurrence have been identified as key factors for limitation of postoperative VA. The observation of progressive RPE atrophy after full macular translocation in patients without apparent recurrent CNV as an additional major factor for loss of VA over time may contribute to long-term prognosis and patient selection. In our study, RPE atrophy appeared late after an operation in eyes without sign for recurrence in angiography and did not include scarring at margins of the lesion as seen after treated or spontaneous transformation of recurrent CNV. Angiographically, these changes were new and different from the original pathology. However, these findings reflect the observations described in patients with geographic atrophy after full macular translocation, in which these changes could be defined as recurrence of the disease at a new location.^17,18 One possible explanation could be that operations in these patients were performed at a stage of the disease when RPE cells and the Bruch membrane under the prospective fovea and the photoreceptors of the original fovea had already been severely damaged and lost the capacity for recovery. Timing of intervention represents a major issue in all therapeutic approaches in the treatment of AMD. Even if involution of the neovascular complex is achieved, significant improvement in VA after treatment, such as photodynamic or intravitreal therapy, may be limited by impaired choroidal perfusion and/or persistent fibrotic lesions remaining beneath the fovea in late-stage AMD.

Complications from a complex operation such as full macular translocation affect functional outcome. Retinal detachment and proliferative vitreoretinopathy are of great concern. In our study, 19% of patients had retinal detachment with proliferative vitreoretinopathy that developed within the first year. An additional 4 patients (4%) had retinal detachment with proliferative vitreoretinopathy in later follow-up. The incidence of retinal detachment and/or proliferative vitreoretinopathy ranges from 17% to 25% in published studies.^9-15 Cystoid macular edema represents one of the most frequent complications, with rates of up to 70%.^9-15 Extent and time of appearance of cystoid macular edema may represent an early sign for struggling RPE at the location of the new fovea. At the final examination in our series, 7 patients showed cystoid macular edema (8%). Recurrent CNV originating from the RPE bed of the removed CNV has been observed in 8% to 21% of patients.^9-15 Reported recurrence rate may be dependent on methods chosen by authors, as sensitivity is critical to detect recurrence and to differentiate it from RPE alteration. Between the 12-month examination and the last follow-up visit, we observed new recurrent CNV, a vascular net with progressive leakage in the late phase, in only 1 patient. Choroidal neovascularization did not recur in any of the 3 patients who had been treated for CNV recurrence during the first 12 months, suggesting that the highest proliferative activity and thus risk for recurrent CNV occurs within the first year after full macular translocation.

The goal of full macular translocation is vision improvement, but there is no evidence to suggest that it does so consistently. Improvement in some patients seems to be offset by deterioration in others. Thus, case selection may be the key to predict outcome and to define the role of full macular translocation.¹⁹ However, data on predictive factors are inconsistent. Mruthyunjaya and coworkers¹⁵ reported that better levels of preoperative visual function were predictive of better levels 12 months after full macular translocation. In our series, the preoperative level of VA did not prove to be a statistically significant factor for change in VA postoperatively. Similar findings have been reported by Wong et al.¹⁹ Lesion size and lesion composition also have been proposed to impact visual outcome. At 12 months, we saw better outcome in patients with hemorrhagic AMD or classic CNV, whereas Wong et al¹⁹ reported poorer functional outcome for patients with minimally classic CNV compared with patients with predominantly classic CNV.¹⁹ In our long-term results, CNV type had no influence on visual outcome. Comparison of results of published studies is limited by difference in follow-up time, inclusion criteria, and surgical and examination techniques. However, there seems to be agreement that early treatment is the key factor to possibly achieve improvement of VA. Compared with alternative approaches of restoring macular RPE function, including membrane extraction alone and membrane extraction combined with iris pigment epithelium transplantation, long-term results after full macular translocation are clearly superior.^5-8 Retinal pigment epithelium transplantation combined with membrane extraction showed improvement of 3 to 8 lines in 38.4% of the cases 12 months after the operation.⁶ These results are similar to those we have obtained after full macular translocation. However, long-term results of RPE transplantation remain to be determined. Given that ours was a single-center study on a small number of patients (coupled with the large-size lesions [median, 2.8 disc diameters] and variable composition of the subfoveal lesion limits), significance of results hinders clinically significant comparison with alternative surgical and nonsurgical interventions in the treatment of AMD.

Even with other treatment options available, full macular translocation may still be considered for selected cases of exudative AMD, such as those who are not eligible or do not respond to alternative treatments. Case selection and decision should carefully consider the risk of early and late systemic and ocular complications.

Correspondence: Sabine Aisenbrey, MD, Department of Ophthalmology, University of Tuebingen, Schleichstrasse 12-16, 72076 Tuebingen, Germany (sabine.aisenbrey@med.uni-tuebingen.de).

Submitted for Publication: September 30, 2006; final revision received April 12, 2007; accepted April 14, 2007.

Financial Disclosure: None reported.

Funding/Support: This study was supported by a grant from the Interdisciplinary Centre for Clinical Research (IZKF) BIOMAT/medical faculty at the Rheinisch-Westfälische Hochschule, Aachen, Germany.