Aggressive posterior retinopathy of prematurity (AP ROP) is characterized by posterior retinopathy usually in zone I, substantial dilatation and tortuosity of the vessels at the posterior pole, a flat network of neovascularization on the retinal surface, circumferential fibrovascular tissue extending for 12 clock hours, and rapid progression to stage 5 without the classic course that includes stages 1 through 3. Previous randomized trials have reported that application of photocoagulation to the nonvascularized retina prevents retinal detachment in classic ROP, but that photocoagulation often cannot prevent progression in eyes with AP ROP.

Because the visual prognosis is poor after vitrectomy performed when AP ROP progresses to stage 5, several early surgical interventions have been tried to prevent stage 5 retinal detachment. Scleral buckling that reduces the traction between the fibrovascular tissue and the retina is ineffective for AP ROP that is characterized by fibrovascular tissue nearly 360 degrees circumferentially in the posterior retina. Lens-sparing vitrectomy usually is successful if performed for tractional retinal detachment in eyes with classic ROP, but the surgery is unsuccessful for AP ROP. Although it is controversial whether lensectomy is necessary during vitrectomy to treat retinal detachment in eyes with AP ROP, wide-field vitrectomy with lensectomy may be required to treat retinal detachment in eyes with AP ROP, which always is characterized by high disease activity despite the disadvantages of lens removal.

Eyes with AP ROP that have undergone early vitrectomy often have a retinal reattachment; however, among these eyes, we recently identified eyes in which proliferation of fibrovascular tissue recurred after surgery. Because a very low birth weight, young gestational age, or long-term high oxygen therapy are risk factors for severe ROP, we analyzed the correlation between recurrent fibrovascular tissue proliferation after early vitrectomy and the associated risk factors.

Methods

The medical records of all eyes that underwent early vitrectomy with lensectomy for stage 4 ROP associated with AP ROP were reviewed retrospectively at the Department of Ophthalmology, National Center for Child Health and Development, Tokyo, Japan, from March 2005 through April 2008. All eyes with AP ROP were diagnosed at the referring hospital or our institution based on the description published in 2005. Eyes with stage 4A or stage 4B ROP in which fibrovascular tissue did not reach the vitreous base were included; eyes with fibrovascular tissue attached extensively to the vitreous base or eyes without dense and early application of photocoagulation even to the nonvascularized retina were excluded, because those eyes usually have a stage 5 retinal detachment, and rigorous evaluation of recurrence of fibrovascular tissue is almost impossible. The follow-up period after vitrectomy exceeded 6 months in all eyes in the analysis. All surgeries were performed by 1 surgeon (N.A.).

Data collected from each case record included gender, gestational age, birth weight, the Apgar score at 5 minutes, the duration of intubation, and the presence of severe systemic complications (i.e., hydrocephalus, patent ductus arteriosus, or necrotizing enterocolitis requiring surgery). We included the duration of intubation as an indicator of the degree of oxygen exposure. ROP findings included the preoperative ROP stage, zone, fibrovascular tissue and vitreous base adhesion, clock hours of fibrovascular tissue, postmenstrual age at the initial application of photocoagulation, photocoagulation to both vascularized and nonvascularized retina ( Figure 1 ) or only to nonvascularized retina ( Figure 2 ), postmenstrual age at vitrectomy, and intraoperative hemorrhage. Although the area of photocoagulation generally is limited to the nonvascularized retina, because the importance of photocoagulation to the vascularized retina has been suggested in eyes with severe ROP, we included eyes treated with photocoagulation applied to both the vascularized and nonvascularized retina. ROP findings were recorded by detailed retinal drawings and RetCam (Massie Research Laboratories, Inc, Pleasanton, California, USA). The eyes that fulfilled the inclusion criteria were divided into 2 groups based on the recurrence or absence of recurrence of fibrovascular tissue after surgery. The factors listed previously were compared between the 2 groups.

Fundus photographs of aggressive posterior retinopathy of prematurity in eyes of patients who underwent early vitreous surgery and preoperative photocoagulation to both the vascularized and nonvascularized retina. (Top left and Bottom left) Fundus photographs obtained before surgery showing densely applied photocoagulation in both the nonvascularized retina and the vascularized retina 3 to 4 spots posterior to the junction. Fibrovascular tissue and a focal tractional retinal detachment are seen at the junction. (Top right and Bottom right) Fundus photographs obtained 4 months after surgery showing that the retinopathy has been treated successfully without recurrence of fibrovascular tissue.

Fundus photographs of aggressive posterior retinopathy of prematurity in the left eye of a patient who underwent early vitreous surgery and preoperative photocoagulation to the nonvascularized retina. (Left) Fundus photograph obtained before surgery showing that photocoagulation was applied to the nonvascularized retina. Reddish fibrovascular tissue and a tractional retinal detachment are observed at the junction. (Right) Fundus photograph obtained 1 month after surgery showing recurrence of fibrovascular tissue in the posterior vascularized retina with a traction retinal detachment.

Statistical analyses were performed using statistical software (StatLab, SPSS for Windows, version 16.0; SPSS, Inc, Chicago, Illinois, USA). Univariate analyses to determine the association between risk factors and recurrence after early vitrectomy were performed using the Mann–Whitney U test, the t test, and the Fisher exact test as appropriate. A multivariate logistic regression model was constructed with recurrence as the dependent variable and with the factors that differed at the significance level of P < .2 in univariate analyses, including the gestational age, preoperative ROP stage, fibrovascular tissue and vitreous base adhesion, the postmenstrual age at the initial application of photocoagulation, and photocoagulation to both the vascularized and nonvascularized retina as independent variables. P < .01 was considered significant.