To analyze risk factors for retinal detachment in eyes with aggressive posterior retinopathy of prematurity (ROP) treated with confluent laser photocoagulation.
Retrospective, consecutive, interventional case series.
Infants undergoing laser treatment for aggressive posterior ROP from January 2006 through June 2010 participated. Eyes with retinal detachment (stage 4A, 4B, and 5) at presentation and those lost to follow-up after treatments were excluded. We defined 2 groups based on outcome after confluent laser photocoagulation: retinal detachment or favorable outcome. Two groups were compared for various factors, which included birth weight, gestational age, postconceptional age at laser treatment, zone of disease, preretinal hemorrhage, pre-existing fibrovascular proliferation, tunica vasculosa lentis, laser spots used for primary treatment, need for repeat laser treatment, and new-onset fibrovascular proliferation after laser treatment. Data were analyzed using univariate and multivariate logistic regression analysis.
One hundred nineteen eyes of 61 infants underwent laser treatment for aggressive posterior ROP. Ten (8.4%) eyes were excluded and 109 eyes (91.6%) were included. Nineteen (17.4%) of 109 eyes progressed to retinal detachment after laser treatment. On univariate analysis, multiple factors were associated with retinal detachment. Of these, all eyes with posterior zone 1 disease and all eyes demonstrating extensive fibrovascular proliferation (> 3 clock hours) after laser treatment progressed to retinal detachment. On multivariate logistic regression analysis, the most significant predictors of retinal detachment were gestational age of less than 29.5 weeks ( P = .006), hemorrhages ( P = .003), need for repeat laser treatment ( P = .006), and new-onset limited fibrovascular traction (< 3 clock hours) after laser treatment ( P = .042).
A gestational age of less than 29.5 weeks, posterior zone 1 disease, and preretinal hemorrhages before laser treatment are the most significant risk factors for retinal detachment in aggressive posterior ROP. Of the events occurring after laser treatment, new-onset fibrovascular traction is associated significantly with development of retinal detachment.
Aggressive posterior retinopathy of prematurity (ROP) is characterized by severe plus disease, flat neovascularization in zone 1 or posterior zone 2, and a rapid progression to retinal detachment in absence of intervention. Various studies have reported a poor outcome in aggressive posterior ROP after laser photocoagulation. Early and aggressive confluent laser photocoagulation has been reported to optimize outcomes in aggressive posterior ROP. However, some eyes with aggressive posterior ROP progress to retinal detachment despite an early, confluent, and adequate laser treatment. Various studies have advocated early vitrectomy for relentless circumferential traction in this scenario. Preoperative dense laser photocoagulation to vascularized and nonvascularized retina has been reported to decrease the recurrence of fibrovascular traction after early vitrectomy. However, little information exists on the predictive factors for a poor outcome after laser photocoagulation in aggressive posterior ROP. We conducted the present study to identify risk factors for retinal detachment in eyes undergoing laser treatment for aggressive posterior ROP.
We performed a retrospective chart review of infants undergoing laser treatment for aggressive posterior ROP at the Advanced Eye Centre, Postgraduate Institute of Medical Education and Research, Chandigarh, India, from January 2006 through June 2010. Aggressive posterior ROP was diagnosed in accordance with a revised international classification of retinopathy of prematurity. ROP was documented by detailed retinal drawings by an experienced retinal surgeon (M.R.D.). Retcam (Clarity MSI, Pleasanton, California, USA) photography was performed wherever possible. All laser treatments were performed using an 810-nm diode laser (IRIS Medical Oculight SL 810-nm infrared laser; Iris Medical Inc, Mountain View, California, USA) by retinal surgeons well versed in ROP treatment (M.R.D., G.S.). Confluent laser burns (less than a half burn width apart) were applied to the entire avascular retina from the flat neovascularization to the ora serrata. Laser treatment also was applied to tongues of clinically avascular retina inside the apparently vascularized posterior retina as suggested by previous reports. Repeat laser treatment to exposed skip areas (after retraction of flat neovascular complex) was carried out at 7 to 10 days after primary treatment. We excluded eyes with retinal detachment at presentation (stage 4A, 4B, and 5). Eyes lost to follow-up after treatments also were excluded. All included eyes completed the 6-month follow-up visit after laser treatment.
Data collected from each case record included birth weight, gestational age at birth, and postmenstrual age at laser treatment. ROP findings included zone of disease, presence of fibrovascular proliferation, preretinal hemorrhages, and tunica vasculosa lentis. For the present study, aggressive posterior ROP was classified into one of the following zones: posterior zone 2, anterior zone 1 (vessels confined to zone 1 but developed beyond the fovea), or posterior zone 1 (vessels posterior to the fovea). We collected data about the number of laser spots during primary treatment; the need for repeat laser treatment and the course after laser treatment, including new-onset fibrovascular proliferation; and progression or regression of disease. Fewer than 3 clock hours of fibrovascular traction was labeled as limited, and more than 3 clock hours was labeled as extensive. All included eyes were classified into 2 groups based on the outcome: favorable or retinal detachment. Favorable outcome meant complete regression without any residual tractional detachment. Retinal detachment meant progression to stage 4A, 4B, or 5 or falciform fold. We compared the above-listed baseline parameters and events occurring after laser treatment for 2 groups to identify factors for retinal detachment.
Univariate analysis to determine the association of various factors with retinal detachment was carried out using the Mann–Whitney U test, independent samples t test, chi-square test, and Fisher exact test, as appropriate. A P value less than .05 was considered statistically significant. The significant factors in univariate analysis were entered into a multivariate logistic regression model with outcome as dependent variable to determine the most significant independent predictors of retinal detachment. A P value less than .05 was considered statistically significant.
One-hundred nineteen eyes of 61 infants underwent laser treatment for aggressive posterior ROP during the study period. Both eyes of 1 infant who had stage 4A ROP at presentation and 8 eyes of 4 infants lost to follow-up after treatment were excluded. Overall, 109 eyes of 56 infants were included. The mean birth weight of included infants was 1392.77 ± 369.55 g and mean gestational age was 30.19 ± 2.55 weeks. The mean postmenstrual age at laser treatment was 35.29 ± 2.2 weeks.
Of the 109 eyes, 90 (82.5%) eyes had a favorable outcome and 19 (17.5%) eyes had an unfavorable outcome. Of the 19 eyes with an unfavorable outcome, 8 (42.1%) had stage 4A ROP, 4 (21.05%) had stage 4B ROP, 4 (21.05%) progressed to stage 5 ROP, and 3 (15.8%) eyes demonstrated falciform fold. On univariate analysis, multiple factors were associated with retinal detachment ( Table 1 ). Presence of posterior zone 2 disease was associated with a favorable outcome ( Table 1 ).
|Favorable Outcome (n = 90 Eyes)||Retinal Detachment (n = 19 Eyes)||P Value|
|Median birth weight (g)||1400||1100||<.001 a|
|Mean gestation ± SD (wks)||30.61 ± 2.54||28.42 ± 1.64||<.001 b|
|Mean postconceptional age at laser treatment ± SD (wks)||35.41 ± 2.21||34.53 ± 2.02||.110 b|
|Posterior zone 2, n (%)||69 (76.7)||1 (5.3)||<.001 c|
|Anterior zone 1, n (%)||21 (33.3)||11 (57.9)||.005 d|
|Posterior zone 1, n (%)||0 (0%)||7 (36.8)||<.001 c|
|Tunica vasculosa lentis, n (%)||18 (20)||11 (57.9)||.001 d|
|Hemorrhages, n (%)||6 (6.7)||6 (31.6)||.006 d|
|Pre-existing fibrovascular proliferation, n (%)||2 (2.2)||4 (21.1)||.008 c|
|Mean no. of laser spots ± SD||2291.59 ± 719.85||3034.95 ± 1370.84||.001 b|
|Relaser, n (%)||7 (7.8)||5 (26.3)||.034 c|
|Limited fibrovascular traction, n (%)||10 (11.1)||6 (31.6)||.033 d|
|Extensive fibrovascular traction, n (%)||0 (0)||13 (68.4)||<.001 c|
All eyes with posterior zone 1 disease demonstrated retinal detachment. All eyes with extensive fibrovascular traction after laser treatment had retinal detachment. Although these factors were highly significant for retinal detachment, they were not entered into the logistic regression model, because they induced error in the overall equation. Receiver operator characteristic curves were constructed for continuous variables including birth weight, gestational age, and laser spots to arrive at cutoff values. Finally, a logistic regression equation was constructed with the following factors, with outcome as dependent variable: birth weight less than 1225 g, gestational age younger than 29.5 weeks, anterior zone 1 disease, tunica vasculosa lentis, preretinal hemorrhages, pre-existing fibrovascular proliferation, more than 2548 laser spots, need for repeat laser treatment, and limited fibrovascular proliferation ( Table 2 ). On logistic regression, the most significant predictors of retinal detachment were gestational age younger than 29.5 weeks, hemorrhages, need for repeat laser treatment, and limited fibrovascular traction after laser treatment ( Table 2 ).