Selective Endothelialectomy in Peters Anomaly: A Novel Surgical Technique and Its Clinical Outcomes in Children : Cornea

Secondary Logo

Journal Logo

Clinical Science

Selective Endothelialectomy in Peters Anomaly: A Novel Surgical Technique and Its Clinical Outcomes in Children

Ramappa, Muralidhar MBBS, MS*,†,‡; Chaurasia, Sunita MBBS, MS*,†,‡; Mohamed, Ashik MBBS, MTech, PhD, MAMS*,§; Ramya Achanta, Divya Sree BS*,†,‡; Mandal, Anil Kumar MBBS, MD, DNB, FAMS; Edward, Deepak Paul MD, FACS, FARVO; Gokhale, Nikhil MBBS, MD, DNB**; Swarup, Rishi MBBS, DOMS, FRCS††; Nischal, Ken K. MD, FAAP, FRCOphth‡‡,§§

Author Information
doi: 10.1097/ICO.0000000000003134


Peters anomaly (PA) is the most common indication for penetrating keratoplasty (PK) among congenital corneal opacities.1 PA is a developmental abnormality in which the corneal defect is mostly limited to the cornea's posterior layers and, if untreated, can result in sensory deprivation amblyopia.1 A step ladder clinical categorization based on clinical signs includes type 1 disease depicted by a central corneal opacity with peripheral iridocorneal adhesions extending from the iris collarette to the perimeter of the corneal opacity (Fig. 1A) and type 2 disease featuring keratolenticular adhesions, which may result in a rudimentary, fragmented, or partially resorbed lens (Fig. 1B).2–4 In Peters plus anomaly, bilateral peninsular-like corneal opacification in the posterior stroma and corresponding defects in Descemet membrane (DM) endothelium complex are noticed besides typical systemic features.5–7 Bilateral disease is seen in 80% of PA,5 and type 2 disease is more often associated with poor visual outcomes than type 1 disease.2,5,8–14 Anterior segment optical coherence tomography (AS-OCT) observation has shown a double-layered DM in PA,15 explaining a reason for spontaneous closure of these defects. Histopathology of PA shows focal abnormal posterior corneal stromal architecture with focal absence of DM (Fig. 1D) and markedly attenuated endothelium with uveal pigment deposits around the defect's edges.16,17

Imaged through a slitlamp and AS-OCT of PA. A, PA type 1 with an posterior excavation involving zone 1. B, Keratolenticular dysgenesis shows failure in separation of lens from the cornea back part (white arrow). C, PA type 1 abnormality with an area of posterior corneal opacification which was shown to be reduplicated DM on pathological specimens (white arrow) at the posterior corneal defect. D, Histological photographs of the cornea showing that Bowman membrane was completely absent in the cornea's middle area. The central internal stroma was hypoplastic, and the fibers were extremely disarranged with few vessels. The Descemet endothelial complex was absent from the cornea (PAS stain, scanner view ×10). E, An abnormally thickened anomalous stratified layer in the DM with pigmented tissues could be found inside the layers (PAS, ×100). F, An abnormal thickened DM with a multiple layering (black arrow) and Descemet endothelial splitting noted (black asterisk) (PAS, ×40).

In the past two decades, the understanding of corneal endothelial cell biology and endothelial transplantation techniques has evolved considerably. For instance, when the peripheral endothelium is healthy, corneal endothelial defects heal rapidly after an injury or insult. However, in PA, the corneal endothelium at the edges and under the defect is abnormal and might inhibit centripetal expansion and normal healing. The absence of central corneal endothelium results in persistent corneal edema and opacification.18 Recent literature supports the idea that the peripheral corneal endothelium is capable of self-renewing after a descemetorhexis in Fuchs endothelial dystrophy.19 Management options in PA include a conservative approach in mild central corneal opacification, with mydriatics and occlusion therapy.20 PA with more extensive central opacities can be surgically managed with a variable success rate. Optical iridectomies,20 rotational keratoplasty,21 conventional full-thickness keratoplasty,2,8,9,11,12,22, and primary keratoprosthesis are the various surgical options used in severe cases.

Visually significant PA treated with PK is highly variable with suboptimal long-term outcomes. Graft failure rates have been reported to be 50% to 100% at 10 years,2 and there is a high incidence of postoperative complications such as allograft rejection, suture-related issues, wound dehiscence, and secondary glaucoma that further compromise the long-term outcomes.2,8,11 Therefore, the surgeon's preference rather than scientific evidence usually determined the technique to be adopted.

Having performed more than 200 full-thickness corneal transplantations in PA with variable results (unpublished), we chose to adopt selective endothelialectomy in Peters anomaly (SEPA), a novel technique for treating corneal opacification in PA in the year 2012. This technique showed that the selective removal of abnormal corneal endothelial cells underlying the posterior defect could reverse corneal edema and promote corneal scar remodeling, eliminating the need for full-thickness keratoplasty. Our initial unplanned experience of spontaneous regression of corneal opacification in keratolenticular dysgenesis after lens removal with endothelial scraping had prompted us to critically review our previous cases that had undergone synechiolysis with endothelialectomy. In addition, we found an anomalous layer over the PA consistently, which was corroborated on AS-OCT (Fig. 1C) and histopathology (Fig. 1D–F). Furthermore, 2 independent studies have published successful outcomes by selective endothelial removal and descemetorhexis.23,24 However, the results needed validation in a large sample with a longer follow-up for broader acceptance of this new technique. Therefore, we report the surgical outcomes of the novel technique SEPA in a large cohort of children with PA.


This was a retrospective, nonconsecutive, interventional clinical case series. A total of 34 eyes of 28 children with nonresolving corneal edema that underwent SEPA at L V Prasad Eye Institute, Hyderabad, India, between 2012 and 2019 were included. The affected eyes of children with PA were evaluated for size and density of the posterior corneal opacity, improvement in visual functions, and AS-OCT of the posterior corneal defect performed before and after surgery in cooperative children. The Institutional Ethics Committee of L V Prasad Eye Institute, Hyderabad, India, approved this study. This study was conducted in strict adherence to the tenets of the Declaration of Helsinki. All legal guardians of children who underwent SEPA and additional interventions (optical iridectomy or endocapsular lens aspiration (LA) with limited anterior vitrectomy) gave written informed consent for all procedures described in this study.

Grading of Disease Severity

We classified the disease severity and zonal involvement as previously defined by Yang et al.2 Clinically, PA is classified based on a constellation of clinical signs. Mild disease was characterized by the presence of a normal iris and lens. Moderate disease was defined by the presence of focal iridocorneal adhesions (anterior synechiae) or other iris defects such as atrophy, abnormal vasculature, or coloboma. Severe disease was defined by the presence of corneolenticular adhesions or corneal staphyloma with or without corneolenticular adhesion.

Zone of Corneal Opacity

The location and extent of the opacity were ascertained based on digital photographs, surgical videos, drawings, and other descriptive information in the medical records. The location of the opacity was described in 3 zones centered on the geographic center of the cornea. Zone 1 (central cornea) was circular with a diameter of 5 mm. Zone 2 (peripheral cornea) extended from zone 1 up to the limbus. Zone 3 (limbal zone) comprised the limbus. The zone of involvement was defined by the most peripheral location of the opacity. Zone 3 involvement was subdivided into subtotal (zone 3A) or total (zone 3B) based on limbal opacification.2

Designation of Laterality

The eyes of patients with visually significant corneal opacity involving both eyes were designated as having significant bilateral disease. The eyes of patients with unilateral disease or asymmetrical bilateral disease in which the corneal opacity in the lesser affected eye was minimal, extraaxial, and visually insignificant were designated as having unilateral disease.

All children who underwent SEPA met the following inclusion criteria: Children with type 1 or 2 PA with a posterior corneal defect limited to zone 3A and affecting visual function and precluding retinoscopy or fundus view under undilated circumstances were considered for SEPA. The eyes excluded were those with pancorneal involvement, zone 3B corneal opacity, anterior staphyloma, secondary glaucoma, microphthalmia, persistent fetal vasculature, aniridia, and those unwilling to undergo a conservative surgical intervention.

Preoperative vision testing, corneal pachymetry, and AS-OCT were performed in all cooperative children. Parents consented with emphasis that this was a conservative surgical intervention that did not need a long-term stringent follow-up and medications and that there was no risk of graft failure. All parents understood that this innovative surgical intervention might have a modest outcome, and definitive intervention such as full-thickness corneal grafting might be necessary in the future. Alternative options including optical iridectomy, rotational autograft, and additional lens removal with concurrent or subsequent PK were discussed with the parents.

Surgical Technique of SEPA

The surgeries were performed under hypotensive anesthesia by experienced corneal surgeons (MR and SC). The anterior chamber was filled with sterile air, and this carefully delineated the actual extent of corneal opacification, any iris strands, and any lens opacity. Trypan blue (Aurovisc, Auro Laboratories Inc, Madurai, India) was used to delineate the extent of corneal opacification and DM reduplication. Using a bimanual irrigation aspiration cannula, the unhealthy corneal endothelium from the opacity's undersurface and at edges was gently scraped off up to 1 mm into normal cornea without disturbing DM's integrity. At the end of the intervention, trypan blue restaining along with a chandelier light pipe was used to identify any focal DM tears or detachment (Video 1,

Additional Interventions

Lysis of visible iris strands was performed when noted to arise from the iris collarette and extend to the perimeter of corneal opacity. Selective endothelialectomy in Peters anomaly with optical iridectomy (SEPAO): Optical iridectomy or sphincterotomy was performed where the corneal opacity was >6 mm or seen obscuring the retinal red reflex without dilation. Optical iridectomy (size of at least 3 clock hours) was performed using a vitrector and aimed to achieve a decent red reflex intraoperatively. Selective endothelialectomy in Peters anomaly with lensectomy (SEPAL): LA along with primary posterior capsulotomy and a limited anterior vitrectomy was performed when the lens was adherent to the posterior corneal surface. All incisional sites were closed using 10-0 nylon sutures. After the procedure, a drop of topical moxifloxacin 0.5% eye drops (Alcon Labs, Fort Worth, TX), prednisolone acetate 1% eye drops (Alcon Labs, Fort Worth, TX), and atropine sulfate 1% eye drops (Aurolab, Madurai, India) were instilled.

Postoperative and Follow-Up Schedule

Postoperatively, all children underwent comprehensive ophthalmic examinations at every follow-up visit either in the outpatient department using the slitlamp when cooperative or in the operating room using a handheld slitlamp under short general anesthesia. The examination included assessing corneal clarity, DM status, lens status, and intraocular pressure (IOP). Visual acuity assessment and refraction were performed when possible. The side ports' sutures were removed 3 to 4 weeks after the surgery. In the postoperative period, the child was maintained on topical moxifloxacin 0.5% eye drops 4 times a day until side ports' sutures were removed. Prednisolone acetate 1% drops 6 times a day were followed by tapering doses of topical steroids over the next 6 weeks. The children were followed up 1, 4, 12, 24, and 52 weeks after the intervention and after that yearly. At each visit, examination under anesthesia was performed for younger children when necessary. If IOP was elevated, topical and, if needed, systemic glaucoma medications were used. Glaucoma filtering surgery was reserved for eyes where IOP was found to be unacceptable on maximal tolerated medical therapy. Retinoscopy was performed at every follow-up visit, either in the outpatient or under anesthesia. Spectacles were prescribed between 1 and 2 weeks after surgery or as soon as corneal clarity allowed a reliable retinoscopy reading. Amblyopia was managed using occlusive therapy, and optical correction with glasses or contact lenses or both was aggressively pursued under the attending surgeon's direction or pediatric ophthalmologist. Intraoperative or postoperative complications were reviewed for all patients. During the first year after surgery, parents were contacted by telephone if they missed a scheduled visit and the next earliest appointment was arranged for them. This was performed proactively to ensure that all patients completed at least 1 year of follow-up.

Optical Coherence Tomography Scan (AS-OCT)

AS-OCT imaging was performed before the intervention either using handheld spectral domain optical coherence tomography (Bioptigen, Inc, Research Triangle Park, NC) under general anesthesia or preoperatively using RTVue-100 Fourier domain optical coherence tomography (Optovue, Fremont, CA) OCT in the office setting in cooperative children. The evaluated parameters included the zone of the extent of corneal opacity, iridocorneal or corneolenticular adhesions, and anterior chamber depth.

Data Collection

Data collected included patient age, sex, laterality, detailed pedigree, details of prior eye surgeries if any, duration of corneal opacification, preoperative visual acuity and IOP, intraoperative surgical details, postoperative complications, length of follow-up, and status of DM, IOP, and optic nerve head health at each visit. Visual acuity was measured using age-appropriate grating acuity in infants and Teller Acuity Cards in young children or linear Sloan letters in older children. Blink or grimace to light was defined as light perception. Centered/steady/maintain or fix and follow were defined as hand motion. Visual acuity values were converted to the logarithm of the minimum angle of resolution (logMAR) equivalence. The significant postoperative primary outcome measures included corneal clarity (cleared, partly cleared, and not cleared). The secondary outcome measures included best-corrected visual acuity (BCVA). The impact of laterality, preoperative age at surgery, size of opacity, additional intervention, and postoperative factors on the clinical resolution of corneal opacification was analyzed.

Outcome Measures of Efficacy

The primary outcome measure was the clinically significant resolution of corneal opacification after SEPA. Successful resolution of corneal opacification was defined as a partial to near-total stromal clearing without epithelial or stromal edema allowing a clear view of iris features, retinoscopy, and fundus evaluation (Figs. 2A–L and 3A–F). Successful SEPA was also described as a reduction in corneal opacification's size and density at the last follow-up without a later surgical procedure. Failure was defined as posterior corneal opacification failing to show clinical resolution even after 3 months after SEPA (Fig. 3D–F), precluding the visibility of iris details, retinoscopy, or retinal examination and a need for additional intervention for visual rehabilitation. The secondary outcome measure of efficacy was the improvement change in BCVA at each postoperative follow-up visit.

Digital photographs showing the dynamics of stromal healing pattern chronologically after SEPA in 2 different children with a mild to moderate disease of type 1 PA. A–F, Right eye of a 6-month-old child (patient 11), the corneal opacity involves the zone 1 with numerous iris strands inserted onto the edge. B, Day 3 postoperatively, after SEPA showing marked corneal edema because of surgical trauma, the same eye 1 month after SEPA showing marked lessening of corneal edema. D, Same eye, 4 months after SEPA was showing a progressive reduction in corneal opacification. E, Same eye, 8 months after SEPA was showing a further decrease in opacity density. F, Last visit, 16 months postintervention, showed a significant reduction in opacity and vision improvement over time. G–L, Captures a similar timeline of events in the left eye of the same child (patient 12).
Digital photographs showing the dynamics of stromal healing pattern chronologically after SEPAL in a child with type 2 PA. A, Right eye of a 7-year-old girl (patient 28) showing corneal opacity involving the zone 1 with keratolenticular adhesion. (B) Day 30 postoperatively, after SEPAL with anterior vitrectomy showed marked improvement visual axis clarity. C, Last visit, 7.6 years postintervention, showed a significant decrease in opacity and vision improvement over time to 20/50. D–F, Captures a similar timeline of events in the left eye of the same child (patient 29) with a severe disease that necessitated SEPAL with an optical iridectomy. Corneal opacity did not regress postoperatively. Finally, the cornea decompensated after Ahmed glaucoma valve implantation.

Cases were categorized based on a clinical resolution of corneal edema and scar remodeling: cleared, partly cleared, and not cleared. Cleared was considered when there was a complete resolution of corneal edema with a discernible scar remodeling by 3 months that allowed a decent retinoscopy reading in undilated condition (Figs. 2A–L and 3A–C). Not cleared had no change in corneal opacity's density and size and where refraction was possible only after complete dilation or through the optical iridectomy opening and was classified as nonresponders (Fig. 3D–F). Partly cleared had a significant clearing but needed pupillary dilation for a decent retinoscopy reading by postoperative month 3 visit (Fig. 3E).

Validation of Diagnosis and Outcome by Independent Masked Evaluators

Five proficient ophthalmologists (2 glaucoma specialists, a pediatric cornea specialist, and 2 cornea surgeons) volunteered as evaluators to validate the investigators' assessment of the type of PA and the outcome of surgical intervention in every case. The evaluators were provided preoperative and postoperative photographs of patients, and they were asked to evaluate the photographs based on the following objective criteria: visual acuity, grading of disease severity, zone of corneal opacity, designation of laterality, and outcome measures of efficacy. The evaluators were masked to the patient's identity and the type of surgical intervention performed. In cases where there was disagreement between the evaluators or surgeons, an independent evaluators' consensus opinion was taken as the final diagnosis or treatment outcome.

Statistical Analysis

The statistical analysis was performed using STATA v14.2 (Stata Corp, College Station, TX). The distribution of continuous data was tested for normality using the Shapiro–Wilk test. Continuous, normally distributed variables were represented as mean ± standard deviation. Median and interquartile range were used to describe continuous nonparametric data. Categorical data were expressed in proportions. A mixed-effects model with a random intercept at the subject level was used to account for the correlation between the fellow eyes of the same patients in the comparison of data between visits and in the analysis of risk factors for failure. In addition, the outcome data were compared among the types of surgical intervention: SEPA alone with/without synechiolysis, SEPAO, and SEPAL. A P value of <0.05 was considered statistically significant.


During the study period (2012–2019), a total of 34 eyes of 28 children diagnosed with posterior corneal opacity due to PA underwent SEPA with or without additional procedures (optical iridectomy or endocapsular LA with limited anterior vitrectomy), as summarized in Supplementary Digital Content 1, Seventeen children (60.7%) were male. Parental consanguinity was noted in 6 of 19 children (31.6%). Nineteen eyes of 13 children with a visually significant bilateral disease accounted for 55.9% of the study cohort, while 15 eyes of 15 children (44.1%) had unilateral involvement.

Table 1 summarizes the baseline demographics by the type of procedure performed. Mean age at presentation, sex, laterality, consanguinity, zone of involvement of the opacity, and preoperative BCVA were comparable among the 3 interventions SEPA, SEPAO, and SEPAL. All eyes with type 2 PA underwent SEPAL.

TABLE 1. - Baseline Characteristics of SEPA, SEPAO, and SEPAL
18 Eyes 7 Eyes 9 Eyes
17 Children 5 Children 6 Children
Age at presentation (yr), mean ± standard error 1.2 ± 0.4 1.0 ± 0.6 0.7 ± 0.5
Sex, male:female 11:7 5:2 2:7
Bilateral presentation, n (%) 8 (44.4%) 4 (57.1%) 7 (77.8%)
Consanguinity 2/11 (18%) 1/6 (16%) 4/6 (66%)
Zone of opacity, n (%)
 Zone 1 8 (44.4%) 2 (28.6%) 2 (22.2%)
 Zone 2 3 (16.7%) 2 (28.6%) 1 (11.1%)
 Zone 3 7 (38.9%) 3 (42.8%) 6 (66.7%)
Severity (type 2 PA), n (%) 0 (0%) 0 (0%) 8 (88.9%)

The mean age at the time of SEPA was 2.06 ± 0.51 years. Twenty-four eyes (70.6%) were in children who were younger than 12 months at first intervention. The postinterventional mean follow-up period was 0.96 ± 0.20 years. Table 2 summarizes the outcomes by the type of procedure performed. SEPA was performed significantly earlier than SEPAO. SEPAO had a significantly shorter follow-up duration than SEPA and SEPAL. SEPAL with limited vitrectomy was performed in 9 eyes. In most of these eyes, the adherent lens was clear (2 eyes), rudimentary (3 eyes), cataractous (3 eyes), or partially resorbed (1 eye). Diligent care was taken to avoid inadvertent lens matter spillage into the vitreous cavity.

TABLE 2. - Postoperative Outcomes After SEPA, SEPAO, and SEPAL
18 Eyes 7 Eyes 9 Eyes
17 Children 5 Children 6 Children
Age at surgery (yr), mean ± SE 1.6 ± 0.6 2.0 ± 0.6 3.4 ± 1.1 <0.0001
Follow-up duration (yr), mean ± SE 0.8 ± 0.2 0.5 ± 0.2 1.8 ± 0.4 <0.0001
Corneal clarity, n (%)
 Cleared 7 (38.9%) 2 (28.6%) 2 (22.2%) 0.85§
 Partly cleared 11 (61.1%) 5 (71.4%) 2 (22.2%) 0.16§
 Not cleared 0 (0%) 0 (0%) 5 (55.6%) <0.0001§
*A P value of <0.01 was considered statistically significant after applying Bonferroni correction for multiple comparisons.
In post hoc analysis by multiple pairwise comparisons, only SEPA was significantly different from SEPAO (P < 0.001) and not SEPA versus SEPAL (P = 0.14) nor SEPAO versus SEPAL (P = 0.23).
In post hoc analysis by multiple pairwise comparisons, SEPAO was significantly different from SEPA (P < 0.001) and SEPAL (P = 0.003) and not SEPA versus SEPAL (P = 0.026).
§Adjusted for age at surgery.
In post hoc analysis by multiple pairwise comparisons, only SEPAL was significantly (P < 0.0001) different from SEPA and SEPAO and not SEPA versus SEPAO (P = 0.98).

Efficacy of Selective Endothelialectomy


Of the 34 eyes, 11 eyes (32.4%) had significant clearing of the corneal opacification. These eyes maintained a clear optical axis with a gradual reduction in the density of corneal haze, which was evident 3 months postintervention, besides expansion of overall corneal diameter, and no further reintervention was necessitated. The proportion of eyes with corneal clearing was comparable among the 3 intervention groups (Table 2).

Partly Cleared

Eighteen eyes (52.9%) had a discernible change in corneal clarity and an acceptable functional vision. Refraction was possible in these patients after pupillary dilation or through an iridectomy opening. In this subset postoperatively, we did not observe any glare or photophobia. The proportion of eyes with partly cleared cornea was also comparable among the 3 intervention groups (Table 2).

Not Cleared

In our surgical cohort, 5 eyes (14.7%) failed to show any discernible resolution in corneal opacification. All these eyes had severe disease involving zone 3A and had undergone SEPAL. In patients (21 and 22) with severe bilateral disease affecting zone 3A, it was found 3 months after the first surgery that the clear optical zone was small, and therefore, optical iridectomy did not give an adequate red reflex. In this patient, the children's parents did not consent for keratoplasty given severe hydronephrosis, but after a repeat large sectoral iridectomy, child had regained decent ambulatory vision. There was a trend toward improved corneal transparency with longer follow-up.

Postoperative Visual Function

Among 34 eyes at presentation, only 1 (2.9%) eye had vision better than 20/190, 20 (58.8%) had an ambulatory vision, and 13 (38.2%) had nonambulatory vision. Postintervention, 24 eyes (70.6%) had ambulatory vision, and 8 (23.5%) achieved a BCVA of 20/190 or better at the final follow-up. Suboptimal visual improvement was due to dense amblyopia and near-total corneal opacification. Case numbers 21 and 22, a patient with bilateral type 2 PA, had a 2-line improvement in visual acuity compared with baseline after a repeat large sectoral iridectomy. The mean spherical equivalent refraction at the final visit was 5.08 ± 1.18 D. The eyes undergoing SEPA and SEPAL had significant improvement in the visual acuity at the final visit when compared with preoperative. In SEPAO, BCVA seemed to improve, but it did not achieve statistical significance (Table 3).

TABLE 3. - Preoperative and Postoperative Visual Functions After SEPA, SEPAO, and SEPAL
Visual Acuity At Presentation Last Visit P *
Number of Eyes (%) Number of Eyes (%)
Overall visual functions (n = 34)
 Nonambulatory 13 (38.2%) 2 (5.9%) <0.001
 Ambulatory 20 (58.8%) 24 (70.6%) 0.27
 20/190–20/50 1 (2.9%) 8 (23.5%) 0.004
Overall: BCVA (logMAR), n = 34, mean ± SE 2.55 ± 0.13 1.78 ± 0.13 <0.001
SEPA: BCVA (logMAR), n = 18, mean ± SE 2.45 ± 0.17 1.70 ± 0.17 <0.001
SEPAO: BCVA (logMAR), n = 7, mean ± SE 2.68 ± 0.15 2.28 ± 0.15 0.051
SEPAL: BCVA (logMAR), n = 9, mean ± SE 2.79 ± 0.19 1.71 ± 0.19 <0.001
*Adjusted for age at surgery and follow-up duration, and a P value <0.007 was considered statistically significant after applying Bonferroni correction for multiple comparisons.
Nonambulatory: hand motions, light motions, light perception or no light perception.
Ambulatory: 20/200 to 20/800, counting fingers, or fixes and follows.

Safety of SEPA in the Operated Eye

None of the eyes that received SEPA and its one of iterations had a DM detachment, persistent corneal edema, lens touch, or uveal trauma. The most common observation in the operated eye was corneal edema at the site of SEPA intervention. The corneal edema resolved spontaneously within 1 month postoperatively, as demonstrated on AS-OCT analysis (Fig. 4A–D). AS-OCT confirmation was available in 23% eyes (8). There were no sight-threatening complications documented. None of the eyes had a decline in visual function postoperatively.

Imaged through a Fourier domain optical coherence tomography of PA. A, Preoperatively, type 1 with a zone 1 PA with a posterior excavation. (B) Postoperative day 1 after SEPA, the posterior layers reflectivity pattern suggestive of intact DM with surrounding stromal edema limited to the surgical site. C, Postoperative 1 week, a well-apposed DM with a marked reduction in the corneal edema. D, Well-apposed DM with a sharp reflectivity pattern suggestive of DM covering the defect 1 month postoperatively.

Five eyes had persistent elevation of IOP after intervention despite excluding eyes with preoperative glaucoma. Four eyes had severe disease, and the other two had a mild–moderate disease. In patients 1 and 2, who had severe disease, but corneal involvement limited to zone 1, have shown a marked improvement in the corneal haze. Ahmed glaucoma valve implantation in the left eye was required to achieve adequate IOP control. Post-Ahmed glaucoma valve intervention, the cornea decompensated, and corneal grafting was deferred in view of dense amblyopia and contralateral aphakic eye that had a BCVA of 6/15 or better (0.40 logMAR) vision. The remaining cases were managed with topical antiglaucoma medications alone.

AS-OCT Analysis

AS-OCT images of preintervention and postintervention were available in 8 eyes (23%). Preoperative image analysis revealed that the Bowman membrane over the corneal defect appeared either disrupted or absent, while the rest of the anterior stromal architecture was comparatively preserved. Overlying the DM defect, the posterior stroma had a denser reflectivity with varying degrees of stromal thinning and excavation. The edges of defects were thickened and rolled out, and few cases showed a presence of anomalous pre-Descemet layer occupying the posterior concavity. There were numerous, delicate iris strands extending between the iris collarette and edges of the corneal defect in most cases. In an example, duplication of DM was observed (Fig. 1C). Type 1 PA (Fig. 4A) corneas that underwent SEPA showed the presence of corneal edema in the immediate postoperative period (Fig. 4B), and an intact DM was seen (Fig. 4C). In summary, AS-OCT analysis suggested a healing pattern after SEPA over the original site of posterior excavation (Fig. 4D).

Resolution of Opacity After SEPA

Table 4 presents associated risk factors for SEPA failure. Using bivariate analysis (P value <0.006), female sex (P = 0.006), disease severity (P < 0.0001), glaucoma (P = 0.001), and additional interventions after SEPA (P = 0.002) were significant risk factors for the failure of SEPA. In multivariate analysis, disease severity (ie, a type 2 PA) was a significant risk factor for the failure of SEPA.

TABLE 4. - Risk Factor Analysis for Failure of Simple Endothelialectomy in Peters Anomaly
Variable SEPA Success (n = 29) SEPA Failure (n = 5) P *
Age at surgery (yr), mean ± standard error 2.07 ± 0.52 2.07 ± 0.51 0.99
Sex, male:female 18:11 0:5 0.004
Laterality, unilateral:bilateral 15:14 0:5 0.02
Severity of PA: Type 1 versus type 2 26:3 0:5 <0.0001
Zone of opacity: 1, 2, or 3A 12:6:11 0:0:5 0.009
Glaucoma, n (%) 3 (10.7%) 3 (60%) 0.001
Microcornea, n (%) 1 (3.5%) 2 (40%) 0.63
Additional intervention, n (%) 3 (10.7%) 3 (60%) 0.002
*A P value <0.006 was considered statistically significant after applying Bonferroni correction for multiple comparisons.

Agreement Between Primary Surgeons and Independent Masked Evaluators

The overall agreement between masked evaluators and primary surgeons for the primary diagnosis and PA type was 100% (95% confidence interval (CI), 87.4%–100% see Supplementary Digital Content 2, The concordance 'kappa' measures between the primary surgeon and masked evaluators for the treatment outcome ranged from 0.51 to 0.61, suggesting they were fair to good.


Our study validates the outcomes of SEPA, a novel and less-invasive surgical strategy to treat corneal opacities in children with PA, and provides surgical outcomes in a large cohort. SEPA is based on the following assumptions in PA: 1) focal area of anomalous DM endothelial complex underlying posterior corneal defect, 2) perilesional healthy endothelium has regenerative potential to fill in the denuded area, 3) robust scar remodeling in children, and 4) relatively well-preserved overlying anterior stromal architecture.17 The anomalous endothelium inhibits normal expansion of the surrounding peripheral corneal endothelium, leading to localized loss of endothelial pump function and corneal edema.18 Therefore, it is logical to believe that there could be an improvement if the abnormal endothelium is removed, resulting in defects being replaced with healthy perilesional corneal endothelium. More recently, the corneal endothelial wound models have demonstrated that after a localized corneal endothelial removal, neighboring cells will enlarge and migrate to restore the anatomic and functional integrity of the corneal endothelial monolayer.19,25,26 Based on these findings, we believe that SEPA, while preserving the DM, might promote spontaneous endothelial migration, thus simply accelerating the spontaneous resolution of corneal edema and opacification, consequently evading limitations PK in children with PA.

Although PA is the most common indication of corneal transplantation among childhood corneal blindness,10 it is challenging to visually rehabilitate these children, considering the complexity of corneal and anterior segment involvement.2,22,27,28 PA type 2 variant often requires sequential lens-based intervention followed by PK.29 Associated comorbidities necessitate further intervention that in turn compromises the long-term graft survival.9 PA type 1 cases typically fare better than PA type 2.2 The visual and functional prognoses in both groups of patients are modest.2 Furthermore, there are several intraoperative challenges and postoperative complications that limit the long-term prospects of PK.2,8 The alternatives to PK include pharmacological dilation, rotational corneal grafts, optical sector iridectomy, and spontaneous resolution, which have had different success rates.20,21,30

In our series, the surgical approach in each case was governed by disease severity, associated comorbidities, child's visual potential, parents' expectations, and surgeon's preference. Several studies give importance to the anatomic and surgical results, and a small subset focuses exclusively on long-term functional outcomes of PK in children with PA. Supplementary Digital Content 3,, provides an overall comparison of this study's results with that of other large series consisting of different surgical interventions for PA.2,8,10–12,14,20,31–34 We realized that PK had an inferior long-term outcome and merited stringent follow-up and had the risk of graft infection, allograft rejection, secondary glaucoma, suture-related issues, astigmatism, graft–host dehiscence, and reintervention if graft fails.11 Because the inclusion criteria and definition of success vary across studies, superficial comparisons sometimes can lead to deeply flawed inferences. Therefore, it may be worthwhile to limit the comparison with our own experience of PK in PA. Our historical results agree with previous studies with an overall modest success rate of 21% at the end of 2 years in children with PA,11 and similar results were mirrored in a recent study published by Sun et al.14

In our surgical cohort, 29 eyes (85.3%; 95% CI, 68.2%–94.5%) that had undergone SEPA with or without additional intervention showed partial to near-total corneal clearing and discernible change posterior to stromal architectures indicative of endothelial cell healing pattern as observed using Fourier domain optical coherence tomography. In 5 eyes of 3 children where cornea failed to show the expected response by 3 months after SEPAL, a larger zone of involvement (Zone 3A) was seen in contrast to eyes that had a successful outcome (zones 1 and 2). A larger zonal involvement of 4 to 6 mm requires more than double the surface area for the remaining endothelium to repopulate (4π vs. 9π).19 By contrast, an 8 mm defect may require a repopulation of 4 times that of the area of a 4 mm of endothelial scraping.19 Therefore, the ideal case selection for this new technique is mild to moderate PA cases with smaller zonal involvement than 7 mm. Although most eyes with successful outcomes showed a marked reduction in corneal edema within 1 month of SEPA, the central corneal haze showed a gradual decline ranging from 6 to 18 months. This suggests that the eyes with denser opacity within the surgical cohort showed the considerable time for stromal clearing and scar remodeling. During this period, children must be monitored closely for amblyopia. If there is no significant improvement in corneal haze based on clinical judgment, consideration should be given to perform an optical iridectomy or PK.

The visual outcomes after SEPA are particularly impressive, and vision continued to improve in 32 eyes (94%) with a longer follow-up duration. Besides, the vision continues to improve with age, owing to visual maturation in growing children. It is known that after SEPA, corneal haze may regress much faster in younger patients because of superior stromal remodeling and better regenerative potential of the peripheral corneal endothelium. Correspondingly, the AS-OCT analysis revealed that a well-preserved anterior stromal architecture helps achieve faster refractive stabilization compared with PK. In addition, corneal diameter expands until the age of 2 years35; thus, a larger optic zone aids the visual rehabilitation. Besides the significant visual gains achieved from this minimally invasive surgical intervention, we have also circumvented long-term postoperative complications associated with keratoplasty, notably the need for stringent postoperative follow-up, the lifetime risk of allograft rejection, and eventual graft failure. To the best of our knowledge, this is the largest report of selective endothelialectomy in the PA being successfully investigated exclusively in children with visually significant corneal opacities.

Outcomes after SEPA are dependent on undisturbed anterior stromal architecture, an island of the transparent cornea in the center or midperipheral region, and healthy peripheral corneal endothelial cells. Endothelial migration requires breakage of contact inhibition of anomalous endothelial cells; hence, adhesiolysis alone would not have allowed cell migration of healthy cells and restoration of endothelial pump function with a lessening in stromal opacity. Although mild–moderate cases where opacity was limited to zone 1 tend to have excellent outcomes, severe cases limited to zone 1 or 2 may fare better if concurrent endocapsular LA with optical iridectomy is considered along with SEPA. Primary PK can be deliberated in cases with extensive corneal involvement (Zone 3B), corneal ectasias, or staphyloma formation after a perforation. Besides more extensive cornea involvement, the visual impairment in PA is frequently due to associated comorbidities; hence, although selective endothelialectomy may improve corneal transparency, the overall visual prognosis may remain poor without other interventions such as prompt amblyopia therapy and glaucoma control.

This study also sheds new light on the possible factors responsible for the failure of SEPA. The analysis revealed that severe disease, presence of microcornea, high IOP, concomitant LA with vitrectomy, and female sex were identified risk factors for nonclearing after SEPA. Inadvertent DM detachment at the time of SEPA should be avoided because it necessitates additional intervention, which in turn adversely affects SEPA outcome. Therefore, it may be advisable to preemptively identify those eyes with wider corneal involvement, corneolenticular attachment, ectasia, extremely thin cornea, perforation, and a staphyloma formation by performing an OCT or ultrasound biomicroscopy before surgery.16,36

This study's major strengths are sample size, single-center design with experienced pediatric corneal surgeons, and long follow-up. Five independent masked evaluators validated the diagnosis, and outcomes were assessed in every case. Other more objective means such as AS-OCT or image analysis of postoperative outcomes could have been used. However, considering that 3/4th of the cohort constituted young children, it would have meant additional anesthesia examination at each follow-up visit to perform AS-OCT or capture images, and these were not feasible. The lack of events (nonclearing) among Type 1 PA suggests that the results may not be generalized to represent the entire cohort of SEPA.

The findings of this study question the paradigm whether this procedure should be offered as a preferred surgical modality in all children with PA or only in select cases, where PK is not the preferred alternative, such as in areas where there is a scarcity of donor tissues, noncompliance, inability to follow-up, or high-risk cases. We have demonstrated, to the best of our knowledge for the first time, that the SEPAL technique can be used even in cases of keratolenticular adhesion as long as the opacity is not greater than 6 mm or at least half of the peripheral cornea is clear. We have shown that, by being used in the manner described by us, SEPA may circumvent these limitations drawing from our favorable experience in infants and children with PA. This study showed that SEPA results are extremely promising, and SEPA can be an effective surgical alternative to optical iridectomy and full-thickness PK benefitting hundreds of children with PA-related blindness worldwide.


1. Kurilec JM, Zaidman GW. Incidence of Peters anomaly and congenital corneal opacities interfering with vision in the United States. Cornea. 2014;33:848–850.
2. Yang LL, Lambert SR, Lynn MJ, et al. Long-term results of corneal graft survival in infants and children with peters anomaly. Ophthalmology. 1999;106:833–848.
3. Chang TC, Reyes-Capo D, Cavuoto KM. Correlation between clinical examination and diagnostic imaging in type II peters anomaly. J Pediatr Ophthalmol Strabismus. 2017;54:395.
4. Osigian CJ, Sayed MS, Kontadakis G, et al. Correlation between age and corneal edema in pediatric patients with Peters anomaly. Int Ophthalmol. 2019;39:2083–2088.
5. Bhandari R, Ferri S, Whittaker B, et al. Peters anomaly: review of the literature. Cornea. 2011;30:939–944.
6. Traboulsi EI, Maumenee IH. Peters' anomaly and associated congenital malformations. Arch Ophthalmol. 1992;110:1739–1742.
7. Nischal KK. Genetics of congenital corneal opacification-impact on diagnosis and treatment. Cornea. 2015;34(Suppl 10):S24–S34.
8. Dana MR, Schaumberg DA, Moyes AL, et al. Corneal transplantation in children with Peters anomaly and mesenchymal dysgenesis. Multicenter Pediatric Keratoplasty Study. Ophthalmology. 1997;104:1580–1586.
9. Yang LL, Lambert SR, Lynn MJ, et al. Surgical management of glaucoma in infants and children with Peters' anomaly: long-term structural and functional outcome. Ophthalmology. 2004;111:112–117.
10. Zaidman GW, Flanagan JK, Furey CC. Long-term visual prognosis in children after corneal transplant surgery for Peters anomaly type I. Am J Ophthalmol. 2007;144:104–108.
11. Rao KV, Fernandes M, Gangopadhyay N, et al. Outcome of penetrating keratoplasty for Peters anomaly. Cornea. 2008;27:749–753.
12. Chang JW, Kim MK, Kim JH, et al. Long-term visual outcomes of penetrating keratoplasty for Peters anomaly. Graefes Arch Clin Exp Ophthalmol. 2013;251:953–958.
13. Yang LL, Lambert SR. Peters' anomaly. A synopsis of surgical management and visual outcome. Ophthalmol Clin North Am. 2001;14:467–477.
14. Sun Y, Lin Q, Miao S, et al. Analysis of graft failure after primary penetrating keratoplasty in children with peters anomaly. Cornea. 2020;39:961–967.
15. Ni W, Wang W, Hong J, et al. A novel histopathologic finding in the Descemet's membrane of a patient with Peters Anomaly: a case-report and literature review. BMC Ophthalmol. 2015;15:139.
16. Nischal KK, Naor J, Jay V, et al. Clinicopathological correlation of congenital corneal opacification using ultrasound biomicroscopy. Br J Ophthalmol. 2002;86:62–69.
17. Townsend WM, Font RL, Zimmerman LE. Congenital corneal leukomas. 2. Histopathologic findings in 19 eyes with central defect in Descemet's membrane. Am J Ophthalmol. 1974;77:192–206.
18. Joyce NC. Cell cycle status in human corneal endothelium. Exp Eye Res. 2005;81:629–638.
19. Borkar DS, Veldman P, Colby KA. Treatment of fuchs endothelial dystrophy by descemet stripping without endothelial keratoplasty. Cornea. 2016;35:1267–1273.
20. Spierer O, Cavuoto KM, Suwannaraj S, et al. Outcome of optical iridectomy in Peters anomaly. Graefes Arch Clin Exp Ophthalmol. 2018;256:1679–1683.
21. Ramappa M, Pehere NK, Murthy SI, et al. Rotational autokeratoplasty in pediatric patients for nonprogressive paracentral corneal scars. Ophthalmology. 2012;119:2458–2462.
22. Yang LL, Lambert SR, Drews-Botsch C, et al. Long-term visual outcome of penetrating keratoplasty in infants and children with Peters anomaly. J AAPOS. 2009;13:175–180.
23. Soh YQ, Mehta JS. Selective endothelial removal for peters anomaly. Cornea. 2018;37:382–385.
24. Hollhumer R, Booysen D. Primary descemetorhexis without graft placement for type 1 Peters anomaly. Can J Ophthalmol. 2019;54:e52–e54.
25. Soh YQ, Peh G, George BL, et al. Predicative factors for corneal endothelial cell migration. Invest Ophthalmol Vis Sci. 2016;57:338–348.
26. Jullienne R, Manoli P, Tiffet T, et al. Corneal endothelium self-healing mathematical model after inadvertent descemetorhexis. J Cataract Refract Surg. 2015;41:2313–2318.
27. Medsinge A, Speedwell L, Nischal KK. Defining success in infant penetrating keratoplasty for developmental corneal opacities. Am Orthopt J. 2014;64:81–88.
28. Zhu AY, Marquezan MC, Kraus CL, et al. Pediatric corneal transplants: review of current practice patterns. Cornea. 2018;37:973–980.
29. Medsinge A, Nischal KK. Cataract surgery in children with congenital keratolenticular adhesion (Peters anomaly type 2). J AAPOS. 2015;19:24–28.
30. Yoshikawa H, Sotozono C, Ikeda Y, et al. Long-term clinical course in eyes with peters anomaly. Cornea. 2017;36:448–451.
31. Junemann A, Gusek GC, Naumann GO. [Optical sector iridectomy: an alternative to perforating keratoplasty in Peters' anomaly]. Klin Monbl Augenheilkd. 1996;209:117–124.
32. Gollamudi SR, Traboulsi EI, Chamon W, et al. Visual outcome after surgery for Peters' anomaly. Ophthalmic Genet. 1994;15:31–35.
33. Frueh BE, Brown SI. Transplantation of congenitally opaque corneas. Br J Ophthalmol. 1997;81:1064–1069.
34. Najjar DM, Christiansen SP, Bothun ED, et al. Strabismus and amblyopia in bilateral Peters anomaly. J AAPOS. 2006;10:193–197.
35. Lagreze WA, Zobor G. A method for noncontact measurement of corneal diameter in children. Am J Ophthalmol. 2007;144:141–142.
36. Yoshikawa H, Ikeda Y, Sotozono C, et al. Ultrasound biomicroscopy in infants with congenital corneal opacity and its correlations with clinical diagnosis and intraocular pressure. Nippon Ganka Gakkai Zasshi. 2015;119:16–21.

corneal scar; optical iridectomy; Peters anomaly; resolution; selective endothelialectomy

Supplemental Digital Content

Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc.