A retinal break following surgery is less forgiving in children than in the elderly because of the difficulty in achieving hyaloid separation1,2 and adequate laser retinopexy during surgery in the presence of an intact hyaloid. In fact, one of the major surgical goals of detachments related to retinopathy of prematurity (ROP) is to relieve traction without creating a retinal break. Spontaneous closure of a retinal break and self-resolution of rhegmatogenous retinal detachment (RRD) is not uncommon in adults,3,4,5,6 but it has not been reported in infants with ROP (based on a Medline search using the terms “rhegmatogenous retinal detachment,” “spontaneous resolution,” “retinopathy of prematurity,” “lens-sparing vitrectomy,” “ROP surgery,” and “retinal detachment after ROP surgery”). In this paper, we report a case of stage 4B ROP that did well with lens-sparing vitrectomy with an attached retina for 4 weeks but developed a retinal break and detachment that resolved spontaneously without any further intervention.
A baby girl born at 27 weeks of gestation with 900 grams birth weight was referred late, at 33 weeks postmenstrual age (PMA). On examination, there was aggressive posterior ROP (APROP) in Zone I. The baby was very sick and was under intensive care in a level II neonatal intensive care unit for 75 days. She had respiratory distress, frequent apneic episodes, poor weight gain, anemia, and thrombocytopenia and was on oxygen supplementation for 40 days. The care provider and the parents did not consent for any intervention including laser or anti-vascular endothelial growth factor. Two weeks later, the retinopathy progressed to stage 4B with iris neovascularization and poor mydriasis. We administered intravitreal bevacizumab (0.625 mg/0.025 mL) (Avastin; Genentech, South San Francisco, CA) at this point of time and topped with laser photocoagulation after 3 days in both eyes. Immediate lens-sparing vitrectomy could not be performed because of lack of fitness. Finally, the surgery was done at 43 weeks of PMA. The uneventful surgery essentially consisted of limited vitrectomy to relieve major tractional vectors; there was no retinal break during surgery (Figure 1). As is our routine practice, we completed the surgery with fluid-air exchange. The air got absorbed within a week and the retina was attached both at 2 weeks and 1 month. However, 2 weeks later, there was a RRD involving the posterior pole with a break and a preretinal membrane in inferotemporal quadrant (Figure 2). The detachment increased for another 4 weeks and a re-surgery was planned. However, at the preoperative evaluation at 53 weeks PMA, the detachment was very shallow and hence we decided to observe. The retinal break, though visible ophthalmoscopically, could not be treated with laser (because of shallow fluid) or cryo (because of its posterior location).
Fundus picture (taken from one of the still frames of a surgical video) of the left eye at the end of lens sparing vitrectomy prior to fluid-air exchange showing an attached retina.
Fundus picture of the left eye showing a retinal break adjacent to a preretinal membrane (blue arrow) and a yellow line demarcating the detached (posterior pole) and attached retina.
In subsequent visits, the retina settled completely in left eye, the hole was no more apparent, and multiple areas of focal depigmentation were noted in the previously detached retina that increased in number and size with time (Figure 3).
Montaged fundus images of the left eye showing sequential changes in the left eye post-vitrectomy. (a) A detached posterior pole with a retinal break in the inferotemporal quadrant; the peripheral lasered retina is attached. (b, c, d, e, f) Spontaneous reattachment of the retina with no apparent hole and multiple depigmented spots at the posterior pole that increased in size and number during follow-up at 2.5, 7, 8, 13, and 18 months post-surgery, respectively.
Appearance of a retinal break and RRD 6 weeks following a successful surgery and finally its spontaneous resolution with post-resolution atypical changes over the previously detached retina was quite interesting in this case.
There was no iatrogenic retinal break, and the retina was attached before fluid-air exchange on operation table and continued for a month after surgery. Fibrosis around the hole could simulate a pseudo-hole. We ruled it out clinically because of the appearance of a choroidal show across the break, increasing retinal detachment during the course of several weeks. Optical coherence tomography (OCT) could have resolved the dilemma, but we did not have an opportunity. We hypothesize that the new retinal break and RRD resulted from the contraction of a trimmed fibrovascular proliferation or traction caused by the adjacent uncut peripheral vitreous cortex. There could have been also an exudative component in addition to the RRD. The time course seems unlikely, though some systemic predisposition to exudate could not be ruled out. We hypothesize four events for retinal hole closure: 1) closed due to a plugging effect by residual uncut vitreous cortex, 2) closure by a transparent membrane bridging the hole, 3) closure by a contracting preretinal membrane, and 4) a spontaneous separation of the preretinal membrane (with subsequent release of surface traction). In the event of break closure, the retinal pigment epithelial pump could have dried the subretinal fluid resulting in retinal reattachment.
There are several reports6,7 on the spontaneous closure of a retinal break and resolution of RRD in nonvitrectomized eyes in elderly. These reports hypothesize the break closure secondary to a combination of release of anteroposterior traction following posterior vitreous separation and plugging effect by vitreous fibers or preretinal membranes. Some of these hypotheses could be applied to our case as ROP surgeries do not involve radical vitrectomy and leaves behind uncut peripheral vitreous cortex. Therefore, to some extent, these eyes behave like nonvitrectomized eyes.
Interestingly we also noted the appearance of multiple areas of focal depigmentation that increased in size and number with time without any pigment clumps in the previously detached retina after spontaneous resolution. This is in contrast to the appearance of demarcation lines, pigmented retinal lesions with convex borders, and subretinal bands in adults.4–7
- Wong SC, Capone A Jr, . Microplasmin (ocriplasmin) in pediatric vitreoretinal surgery: Update and review. Retina. 2013;33(2):339–348. doi:10.1097/IAE.0b013e31826e86e0 [CrossRef]
- Sebag J. Age-related differences in the human vitreoretinal interface. Arch Ophthalmol. 1991;109(7):966–971. doi:10.1001/archopht.1991.01080070078039 [CrossRef]
- Cantrill HL. Spontaneous retinal reattachment. Retina. 1981;1:216–219. doi:10.1097/00006982-198101030-00016 [CrossRef]
- De Juan E Jr., Machemer R. Spontaneous reattachment of the retina despite proliferative vitreoretinopathy. Am J Ophthalmol. 1984;97(4):428–433. doi:10.1016/S0002-9394(14)76125-2 [CrossRef]
- Kokolakis SN, Bravo L, Chignell AH. Late retinal reattachment. Br J Ophthalmol. 1981;65(2):142–146. doi:10.1136/bjo.65.2.142 [CrossRef]
- Cho HY, Chung SE, Kim JI, Park KH, Kim SK, Kang SW. Spontaneous reattachment of rhegmatogenous retinal detachment. Ophthalmology. 2007;114(3):581–586. doi:10.1016/j.ophtha.2006.05.080 [CrossRef]
- Chung SE, Kang SW, Yi CH. A developmental mechanism of spontaneous reattachment of rhegmatogenous retinal detachment. Korean J Ophthalmol. 2012;26(2):135–138. doi:10.3341/kjo.2012.26.2.135 [CrossRef]