Persistent fetal vasculature (PFV), previously termed persistent hyperplastic primary vitreous (PHPV) by Algernon Reese in the 1940s, is one of the most common congenital malformations affecting the eye.1 The hyaloid artery fails to regress in newborn eyes with PFV, resulting in vascular remnants and anatomic abnormalities, including persistent pupillary membranes and posterior tunica vasculosa lentis.2 More than 50% of PFV patients are considered to have unilateral disease;3 however, unilaterality has generally been established by indirect ophthalmoscopy and fundus photography in the past.
Recently, fluorescein angiography (FA) using a wide-field imaging system has been more widely adopted; however, only limited research has been conducted to identify fluorescein angiographic findings in contralateral eyes of patients with unilateral PFV. In this study, we aimed to use this newer technology to identify vascular abnormalities in the contralateral eye of patients diagnosed with unilateral PFV.
Patients and Methods
This retrospective study was conducted at a tertiary-level pediatric ophthalmology center at Xin Hua Hospital, affiliated to Shanghai Jiao Tong University School of Medicine. Consecutive patients who were diagnosed with unilateral PFV and who underwent FA examinations from August 2014 to January 2019 were included. Complete medical records were reviewed, and patients who were followed for fewer than 6 months were excluded. Additional exclusion criteria for this study were as follows: patients with a family history of ocular diseases; patients with a history of prematurity; patients with clinically bilateral involvement; patients with systemic symptoms.
This study was conducted with Institutional Review Board approval and adhered to the tenets of the Declaration of Helsinki. The legal guardian of each patient signed a consent form before all examinations and treatments.
Fluorescein Angiography and Description
FA was performed by one experienced doctor (QZ) in all patients using the RetCam3 system (Natus Medical Incorporated, Pleasanton, CA), with a standard pediatric dose (7.7 mg/kg body weight) of intravenous fluorescein sodium (Fluorescite; Alcon, Fort Worth, TX) under general anesthesia. For patients with complications of PFV requiring intraocular surgery, including cataracts or retinal detachment, FA was conducted after surgery and only in the contralateral eye to avoid infection. For patients undergoing noninvasive procedures such as an examination under anesthesia (EUA) or laser indirect ophthalmoscopy (LIO), FA was conducted before or after those procedures in both eyes. During FA, scleral indentation was utilized to image the ora serrata region if necessary; however, those photos were not used for measuring the width of peripheral avascular retina to avoid measuring deviations caused by ocular deformation (See video below).
The primary outcome of the study was the width of the temporal avascular area between the edge of the vessels and the ora serrata. The secondary outcome was the presence of any peripheral vascular abnormalities in the contralateral eyes of patients with unilateral PFV based on FA images. The width from the retinal vessel termini to the ora serrata temporally was measured and divided by disc diameters (DD) using the “straight tool” of ImageJ 1.52a (Wayne Rasband, National Institutes of Health, Bethesda, MD) on MacBook (Apple, Cupertino, CA). If vascular termini were visible but the ora serrata was not, the width of the peripheral avascular retina was measured from the vessel termini to the edge of the image, thus underestimating the actual width.4,5
Statistical Analysis and Photo Editing
All analyses were performed using the Statistical Package for the Social Science software version 22 (SPSS Inc., Chicago, IL). Kolmogorov-Smirnov tests were used to analyze the distribution of samples. Independent t-tests and analysis of variance tests were performed to compare means between groups. A P value of less than .05 was considered statistically significant. Brightness and contrast of FA images were enhanced by Microsoft PowerPoint for Mac Version 16.37 (Microsoft Corporation, Redmond, WA), and figure formats were converted to TIFF files by Photo Image Editor Pixelstyle Version 3.6.5 (EffectMatrix Inc., Little Neck, NY).
Forty-five patients (34 males) were included in this retrospective study. The mean age at FA performance was 17.31 ± 16.86 months (range: 1 to 74 months). PFV affected the right eye in 25 patients and the left eye in 20 patients. Eleven patients had anterior PFV, two patients had posterior PFV, and 32 patients had combined PFV. Five patients had FA performed in both eyes, and 40 patients had FA performed only in the contralateral eye.
Temporal Peripheral Avascular Area in Contralateral Eyes
The mean width of the temporal peripheral avascular area in contralateral eyes was 1.87 ± 0.71 DD, with a median of 1.80 DD. Sixteen patients (35.6%) had temporal peripheral avascular areas of 2 DD or greater in the contralateral eye; 27 patients (60%) had avascular areas less than 2 DD but equal to or greater than 1 DD, and two patients (4.4%) had avascular areas less than 1 DD (Figure 1). The ora serrata was not visible in the FA images of 37 patients (82.2%); accordingly, the width of the peripheral avascular retina was measured from the vessel termini to the edge of the image in these patients, as mentioned above. There were no statistical differences in the mean widths of the peripheral avascular areas in different types of PFV.
Fluorescein angiogram photos showing temporal peripheral avascular areas in the contralateral eyes of three patients with unilateral persistent fetal vasculature. (A) The temporal peripheral avascular area was less than 1 disc diameter (DD) in patient A (0.56 DD). (B) The avascular area was less than 2 DD but greater than 1 DD in patient B (1.4 DD). (C) The avascular area was greater than 2 DD in patient C (2.4 DD).
Peripheral Vascular Abnormalities in Contralateral Eyes
Vessel shunts, vascular tortuosity and dilatation, circumferential vessels, and abnormal capillary beds in the peripheral retina were observed in 23 (51.1%) contralateral eyes of patients with PFV (Figure 2). Among these patients, seven had anterior PFV, one had posterior PFV, and 15 had combined PFV.
(A–F) Fluorescein angiograms demonstrating vascular abnormalities in the contralateral eyes of five patients with unilateral persistent fetal vasculature (E and F are from the same patient). Vessel shunts along vascular-avascular junctions (arrows); vascular dilatation and staining (open arrows); vascular tortuosity (open arrowhead); circumferential vessels with radial branches (closed arrowheads); and abnormal lacy or feathery capillary beds (observed most clearly in the boxed areas).
All patients with abnormal contralateral eyes were managed by observation. During the follow-up period (mean: 28.97 ± 14.84 months, range: 6 to 61 months), peripheral avascular areas and vascular anomalies remained stable, according to fundus photos obtained by RetCam.
Symmetry of Angiographic Patterns in Both Eyes
In one patient, we observed abnormal vascular shunting and hyperfluorescence in the temporal retina in both his PFV eye and his contralateral eye. In another patient, we observed similar hook-like vascular patterns in both eyes (Figure 3).
Fluorescein angiogram (FA) photos of patients with persistent fetal vasculature (PFV). (A) FA photos from a 9-month-old boy with PFV in his right eye and (A, B) abnormal vessel shunts along the vascular-avascular junctions (arrows) and hyperfluorescence in the temporal retina observed in both eyes. (C) FA photos from an 11-month-old boy with PFV in his right eye and (C, D) abnormal hook-like vascular patterns and vessel shunts (arrows) observed in both eyes.
In this study, we used FA to observe retinal vascular features in the contralateral eyes of patients with unilateral PFV and measured the width of the temporal peripheral retinal avascular area. We found that more than 90% of patients had a temporal peripheral avascular retina of 1 DD or greater in the contralateral eye, and approximately 40% of patients had a temporal peripheral avascular retina of 2 DD or greater. In addition, 42% of patients with PFV had vascular abnormalities in the contralateral eyes, including vessel shunts, vascular tortuosity and dilatation, circumferential vessels, and abnormal capillary beds.
To our knowledge, limited research has previously been conducted to investigate the fluorescein angiographic findings in the contralateral eyes of patients with PFV,6 and the present study had the largest sample size of these related studies. FA images have, however, recently been used to show nonperfused peripheral retinas in the contralateral eyes of patients with Coats' disease.5,7 Although Coats' disease has generally been considered to be more frequently unilateral, those study findings may suggest otherwise. Similarly, the present study suggests that PFV may be also be a bilateral disease with asymmetry in presentation.
Incomplete peripheral vascularization is common in some pediatric retinal diseases, especially familial exudative vitreoretinopathy (FEVR) and retinopathy of prematurity (ROP). Additionally, incomplete peripheral vascularization has been observed even in healthy adults. Using ophthalmoscopy, Rutnin and Schepens showed that it is typical for the peripheral fundus to have approximately 0.5 DD of peripheral retinal nonperfusion.8 Asdourian and Goldberg reported approximately 1 DD of nonperfusion in young healthy adults.9 Blair et al. used FA to evaluate the length of retinal nonperfusion in 33 pediatric eyes with clinically normal-appearing periphery, and found it to be 1.5 DD or less, with a mean of 0.9 DD temporally.4 Those 33 patients had ocular diseases either only in the contralateral eye or, if in the study eye, to a degree judged by the authors not likely to affect peripheral retinal vascular development. Among those patients, four had PFV in the contralateral eye.4 Blair et al. concluded that 2 DD or greater of nonperfusion, or 3 standard deviations above normal, should be considered abnormal and a sign of peripheral nonperfusion.4
In our study, 35.6% of patients had peripheral temporal avascular retina of greater than or equal to 2 DD; however, no complications related to these avascular retinas were observed during follow-up. Accordingly, we would not yet consider the peripheral avascular areas to be abnormal in these patients. There is evidence, however, that peripheral avascular retinas are associated with an increased risk of retinal breaks and extraretinal fibrovascular proliferation in children with ROP and FEVR, as well as longer axial lengths and lattice degeneration leading to retinal detachment in adults.10–12 This study's patients will undergo long-term follow-up to monitor for the development of retinal holes, fibrovascular proliferation, and retinal detachment and to determine if ocular function is affected.
In this study, vascular abnormalities, including vessel shunts, vascular tortuosity and dilatation, circumferential vessels, and abnormal capillary beds, were identified in the contralateral eyes of patients with unilateral PFV. These findings are reminiscent of ROP, a neovascular retinal disorder resulting from delayed or abnormal retinal vessel growth in premature infants,13,14 with similar peripheral vascular patterns noted in ROP patients after bevacizumab (Avastin; Genentech, South San Francisco, CA) treatment.13,14 The present study's findings indicate that abnormal retinal vessel growth may also occur in the contralateral eyes of patients with PFV. Based on these findings, FA is highly recommended during EUAs, not only for children with bilateral diseases like FEVR or incontinentia pigmenti (IP), but also in children with presumed unilateral diseases such as Coats' disease or PFV. We also suggest modifications to the terminology used to describe PFV, with the clinically affected eye being referred to as “abnormal” or in the “clinical stage” and the contralateral eye being referred to as “subnormal” or in the “subclinical stage.”
The underlying mechanisms for the peripheral avascular retina and the abnormal vasculature in our cases is unknown. In truth, the exact regulatory mechanisms responsible for PFV even remain poorly understood. Several gene knockout (KO)/transgenic mouse models have been shown to manifest features of PFV, including the βA3/A1-crystallin model, Ephrin-A5 KO model, Lama1 KO model, Collagen type XVIII and type XV KO models, Atoh7 mutant mice, Frizzled 5 KO model, and Bax and Bak KO models.15
In addition, a case series has been reported regarding the phenotypic overlap between FEVR and PFV associated with the Frizzled 4 mutation.16 Abnormal retinal vasculature has also been observed in some gene KO mouse models. It is possible that the gene mutations that result in the cellular defects associated with PFV may also be involved in the formation of abnormal retinal vasculogenesis and angiogenesis.
The limitations of this study included its retrospective nature, referral bias, admission bias, small sample size, and the lack of a control group. Larger sample sizes and longer follow-up periods are needed to corroborate our findings and to observe long-term complications in these patients.
In conclusion, vascular abnormalities are more common in the contralateral eyes of patients with unilateral PFV than previously reported. Patients with PFV should undergo careful evaluations of both eyes, preferably with fluorescein angiography, to identify vascular changes not visible clinically.