Retinal Hemorrhages During Examination for Retinopathy of Prematurity: A Form of Ocular Decompression Retinopathy

Abstract

Purpose:

Posterior segment hemorrhage occurring during or shortly after examination (PSHE) for retinopathy of prematurity (ROP) is a very rare complication. We present a case of and review the literature on PSHE during ROP examination to better characterize this complication.

Methods:

A case report is presented, followed by a review of similar cases in the literature.

Results:

An infant undergoing laser photocoagulation for ROP rapidly developed diffuse intraretinal hemorrhages in his right eye during the laser and after a Valsalva event while he was intubated under general anesthesia. The hemorrhages resolved within 1 week. This presentation was similar to those in previously reported cases.

Conclusions:

PSHE in ROP usually consists of multiple, diffuse, intraretinal hemorrhages that occur within minutes of ROP examination and resolve within a few weeks without any other ocular findings or sequelae. PSHE seems to represent a form of ocular decompression retinopathy.

Introduction

Posterior segment hemorrhage (PSH) of one or more types, including vitreous, preretinal, intraretinal, subretinal, and choroidal hemorrhage, is a well-known complication of laser photocoagulation (LPC) and intravitreal antivascular endothelial growth factor injection (IVI) for retinopathy of prematurity (ROP). ^1,2 However, PSH occurring during or shortly after examination alone (PSHE) is a very rare complication of ROP, generally described only in case reports as multiple hemorrhages (usually intraretinal, but also preretinal and subretinal) throughout the vascularized retina with rapid onset and resolution within weeks. ^{3 -10} Because the fundus appearance of PSHE mimics that of other pediatric hemorrhagic retinopathies, most notably nonaccidental trauma, it is important to recognize PSHE. However, to our knowledge, the PSHE syndrome has not yet been comprehensively characterized. We describe another case of PSHE, review the literature on PSHE, and present a unifying characterization of this condition.

Methods

The MEDLINE and Embase databases were each searched by the first author in April 2020 for all English-language articles reporting PSH found during or within 1 week of examination for ROP. The 1-week limit was chosen to help exclude any hemorrhages occurring because of evolution of retinal neovascularization. The following searches were performed: “hemorrhage AND retinopathy of prematurity,” “hemorrhage AND scleral depression,” and “hemorrhage AND scleral indentation.” Only articles detailing the temporal association between the ROP examination and the PSH, and without a clear cause of PSH unrelated to the examination, were included in our analysis. Because ROP examination usually happens during LPC and IVI, articles noting PSH during or shortly after these procedures for ROP were also included as long as the PSH was not due to the treatment directly. The fundus photos of PSHE provided in these articles were all evaluated by the first author for aspects of the hemorrhages not provided in the texts. Our case report and the previously published cases were descriptively analyzed together.

Results

Case Report

A male infant born at 24 weeks’ gestational age (GA) with a birth weight of 586 g underwent bilateral intravitreal ranibizumab injections at 33 weeks’ and 41 weeks’ GA for initial and recurrent type 1 ROP, respectively. His comorbidities included aortic root dilation, necrotizing enterocolitis, osteopenia of prematurity, and anemia of prematurity. During outpatient follow-up ROP exams, he was scheduled for LPC for a persistently avascular retina. At 57 weeks’ GA, the patient underwent a dilated fundus exam, RetCam fundus photos and fluorescein angiography (FA; Natus Medical, Inc), and LPC (with scleral depression) to the avascular retina, all under general endotracheal anesthesia. He was found to have zone 3, stage 0, no-plus ROP (Figure 1). The LPC was completed uneventfully in the right eye, but a few minutes after the LPC was started in the left eye, the infant coughed several times against the endotracheal tube. A few minutes later, once the anesthesiologist adjusted the infant’s sedation, both eyes were re-examined. The patient was found to have diffuse intraretinal flame and dot-blot hemorrhages throughout only the right vascularized retina, more densely spaced in the posterior pole and along the border of retinal vascularization (Figure 2). There was no optic disc edema, and the left eye remained unchanged. All hemorrhages resolved by 1 week postoperatively. Because of this rapid resolution without any obvious ocular sequelae, no workup was pursued for bleeding diathesis or other contributory factors.

Figure 1.

Color photos and fluorescein angiography show stage 0 retinopathy of prematurity in each eye.

Figure 2.

Color photos show diffuse flame and dot-blot hemorrhages concentrated in the posterior pole and along the edge of retinal vascularization.

Literature Review and Synthesis of Case Data

The search query described earlier yielded only 8 case reports that met the inclusion criteria and were included for analysis. Tables 1 and 2 list the features and comorbidities of the PSHE patients, respectively, and Table 3 details the features of the hemorrhages.

Table 1.

Characteristics of Patients With Posterior Segment Hemorrhages During Examination for Retinopathy of Prematurity.

First author, year	Birth GA, wk	Birth weight, g	GA at time of hemorrhages	Zone and stage of ROP at time of hemorrhages
Adams, 2004 3	25	920	32	S0 (zone U)
Lim, 2006 4	U	U	45	Z3, S0
Mavrofrides, 2006 5	24	635	40	Z2, S3
Jensen, 2011 6	26	850	38	Z2, S3, pre-plus
Phalak, 2014 7	35	1600	39	Z2, S3, pre-plus
Chandra, 2016 8	32	1212	44	Z3, S0
Yusuf, 2017 9	24	750	38	Z2, S0
Hussain, 2019 10	25	718	35	Z1, S3
Matei, 2020	24	586	57	Z3, S0

Abbreviations: GA, gestational age; ROP, retinopathy of prematurity; S, stage; U, unreported; Z, zone.

Table 2.

Systemic Comorbidities of Infants With Posterior Segment Hemorrhages During Examination for Retinopathy of Prematurity.

Systemic comorbidity	No. of patients
Respiratory failure or other lung disease	4
Anemia of prematurity	2
Necrotizing enterocolitis	2
Neonatal jaundice	2
Osteopenia of prematurity or other bone disease	2
Patent ductus arteriosus (status postligation)	2
Seizures or seizure-like activity	2
Aortic root dilation	1
Apnea of prematurity	1
Acquired protein C deficiencya	1
Sepsis	1

^a Of 5 patients for whom coagulopathy labs were checked, ^3,4,6,7,9 only one patient, ⁷ who had sepsis, had an abnormal result: acquired protein C deficiency, which is known to occur in sepsis, ¹¹ and that spontaneously resolved within 6 months.

Table 3.

Characteristics of Posterior Segment Hemorrhages During Examination for Retinopathy of Prematurity.

First author, year	Type of hemorrhages	Laterality	Location	Time to discovery	Time to resolution
Adams, 2004 3	Flame, dot-blot	Left eyec	Diffuse	Immediate (30 min)	U; nearly resolved by 2 wks
Lim, 2006 4	Flame, dot-blot	Right eye	Posterior pole, peripapillary	Immediate (10 min)	14 wks
Mavrofrides, 2006 5	Subretinal (round)a	Right eye	Diffuse		2 wks
Jensen, 2011 6	Flame, dot-blot, white-centered	Left eye, then right eye 1 d later	Diffuse	1 wkd	U; starting to improve at 4 d
Phalak, 2014 7	Flame, dot-blot, white-centeredb	Left eye	Diffuse	Immediate	6 mo, but improved at 2 wks
Chandra, 2016 8	Flame, dot-blot, white-centered	Right eye	Diffuse, to the equator	Immediate (15 sec)	4 wks
Yusuf, 2017 9	Flame, dot-blot, white-centered	Right eye	Diffuse, denser posteriorly and near vascular border	Immediate	1 wk
Hussain, 2019 10	Confluent, large blots	Left eye	Circumferential hemorrhages only near vascular border	Immediate	2 wks
Matei, 2020	Flame, dot-blot	Right eye	Diffuse, denser along vascular border	Immediate	1 wk

Abbreviations: PSHE, posterior segment hemorrhages during examination; ROP, retinopathy of prematurity; U, unknown.

^a Retinal vessels were seen crossing these hemorrhages, but because they may also cross over deep intraretinal hemorrhages, and because no ultrasonography or optical coherence tomography was performed, these hemorrhages could have instead been primarily intraretinal.

^b This patient was thought to have experienced a central retinal vein occlusion. However, the fundus photos did not show optic disc edema or tortuous, dilated veins, and most of the hemorrhages resolved within 2 weeks. So, this case could have been merely a more severe case of PSHE.

^c PSHE happened a second time in the same eye during a follow-up examination 2 weeks later.

^d The fundus had not been examined in the interim, so the hemorrhages could still have occurred immediately after the examination; PSHE was found in the fellow eye 1 day after a subsequent examination.

At the time of the PSHE, the patients’ mean GA was 41 weeks (range, 32-57 weeks). Although none of their retinas were fully vascularized, all had zone 2 or greater vascular maturation at the time of PSHE, and 5 of 9 patients (56%) had stage 0 ROP. All 9 patients underwent either a scleral-depressed dilated fundus exam or (in 2 patients) RetCam photography shortly before the occurrence of hemorrhages. Only 3 of the patients (33%) also underwent otherwise uncomplicated LPC during the same encounter. Two patients (22%) had a documented Valsalva event during the examination, due to crying ⁸ or (in our case) coughing against the endotracheal tube while under anesthesia. The infants otherwise had no acute distress, vital sign instability, or bleeding from any other sites during the examination or treatment.

In all cases, PSHE occurred in only one eye per examination session, nearly equally distributed between the right and left eyes. In 8 of 9 (89%) patients, the PSHE was noted during that encounter. The PSHE consisted of multiple intraretinal hemorrhages (usually less than 1 disc diameter in size) in at least 8 of 9 (89%) patients, including dot-blot, flame, and white-centered in 8 of 9 (89%), 7 of 9 (78%), and 4 of 9 (44%) of patients, respectively. In 3 patients, ^{7 -9} the fundus photos also suggested the presence of preretinal hemorrhages near the border of retinal vascularization, and in one patient ⁶ the hemorrhages were reported to be solely subretinal.

The hemorrhages were diffuse throughout the vascularized retina in 7 of 9 (78%) cases. In our patient and 3 others, ^6,7,9 the hemorrhages appeared more densely distributed in the posterior pole, as well as adjacent to the border of vascularized retina. FA, performed after PSHE in 2 patients, ^4,9 showed only blocking defects corresponding to the hemorrhages. PSHE resolved completely in a median of 2 weeks (range, 1-24 weeks), without any known ocular sequelae.

Conclusions

From these data, we can construct a syndrome for PSHE in ROP: an acute, unilateral onset of multiple fundus hemorrhages, usually intraretinal and occasionally preretinal or subretinal. The hemorrhages are diffuse and more densely concentrated in the posterior pole and near the border of vascularized retina; in a minority of cases, they may be confined to either the former or the latter. Optic disc edema, changes in vascular caliber, and changes in posterior retinal perfusion are usually absent on FA. The hemorrhages usually appear within minutes of the examination, and they improve substantially or resolve completely within a few weeks without any notable long-term effects. Contrary to previous reports, ⁶ PSHE can happen regardless of ROP stage.

The suspected precipitating factor common to all the PSHE cases is application of pressure to the globe during the ROP examination. We hypothesize that PSHE is due to capillary damage from high intravascular pressure transmitted to them as a result of a lag in autoregulatory compensation following a sudden decrease in intraocular pressure (IOP). Ocular perfusion pressure (OPP) is the difference between mean central retinal arterial pressure and IOP, and retinal autoregulation is thought to be mediated myogenically and metabolically, mainly at the arteriolar level, within seconds of OPP changes. ¹²

Forceful scleral indentation may raise IOP to above central retinal arterial perfusion pressure, and the ensuing retinal metabolic changes, including hypoxia and hypercapnia, stimulate arteriolar vasodilation. Then, when the scleral depressor is released from the globe, the IOP suddenly decreases and the suddenly high intravascular pressure gets transmitted to the retinal capillaries, while the arterioles are still dilated and awaiting autoregulatory vasoconstriction in response to increased blood flow. The overwhelmingly high intravascular pressure in the capillaries likely damages their walls, causing extravasation of blood (flame and dot-blot hemorrhages). This mechanism is amplified by more forceful indentation (a higher peak-baseline IOP gradient), more rapid release of indentation (a faster drop in IOP), and repeated globe indentation and release during the same encounter (repeated cycles of damage to the capillary wall).

The underdeveloped vessels and autoregulatory capacity in infants with ROP may explain why PSHE occurs in this population. ¹³ Additional circumstances may also contribute to the occurrence of PSHE. One such factor is a concurrent Valsalva maneuver, which may further contribute to high capillary intravascular pressure by a suddenly elevated central venous pressure transmitted via the venules. However, a Valsalva event in the setting of PSHE may not result in preretinal hemorrhages, as in classic Valsalva retinopathy; our case of PSHE following a Valsalva event featured only intraretinal hemorrhages. Another factor that may contribute to PSHE is a comorbid blood dyscrasia. PSH following LPC for ROP has been reported in the setting of thrombocytopenia. ¹⁴

Overall, the key pathophysiologic factor underlying PSHE may be the sudden and large drop in IOP, which is also the key factor underlying ocular decompression retinopathy (ODR). ODR is a hemorrhagic retinopathy associated with rapid and large decreases in IOP that was initially described after glaucoma filtering surgery but has since been noted after a variety of procedures. ¹⁵ PSHE is clinically highly similar to ODR. Like PSHE, ODR involves a rapid onset of PSH, most commonly flame, dot-blot, and white-centered retinal hemorrhages, either intraoperatively or by the following day in the majority of cases. As in PSHE, the hemorrhages in ODR are diffusely distributed, more densely in the posterior pole. FA in ODR, as in PSHE, merely shows blocking defects due to the hemorrhages. Finally, the hemorrhages in ODR, like those in PSHE, usually resolve within several weeks without any identifiable ocular sequelae.

Hence, we propose that PSHE in ROP is a form of ODR and should be referred to as ODR. We recommend that the pressure of globe indentation during ROP exams be minimized, that globe indentation be slowly released, and that repeated scleral indentation of the same eye(s) be minimized during the same encounter. In addition, a systemic workup for a bleeding diathesis should be considered either in anticipation of or following PSHE.