Pediatric penetrating keratoplasty (PK) has been recognized as high-risk keratoplasty because of difficulties in the surgical procedure and postoperative management. PK in pediatric patients leads to poor surgical outcomes compared with those obtained in adults, and congenital corneal opacities (CCOs) show a lower survival rate than those of acquired corneal diseases after pediatric PK.1 However, PK is the first-line treatment for CCOs, especially Peters anomaly (PA), in children.2 Scholars have reported graft survival after PK in children with CCOs range from 32.6% to 78.6%.1,3–10 Younger children seem to exhibit worse graft survival.1
However, the definition of successful corneal transplantation in children is not only graft survival: vision rehabilitation is also very important. The period of 0 to 3 years after birth is critical for visual development.11 Early PK is necessary to reconstruct visual function and avoid amblyopia for children with CCOs.10 Nevertheless, we found that not all children obtained optimal visual acuity even though they achieved transparent grafts. In our previous study,12 50 children aged 0 to 7 years with CCOs obtained clear grafts after PK. Vision assessment showed that only 23.3% achieved visual acuity of 20/260. However, the mean age at PK of these children was 2.5 years, and they did not receive aggressive therapy against amblyopia. Whether the unsatisfactory vision in transparent grafts was associated with the late age at which primary PK was performed, or due to neglected treatment for amblyopia, is not known. Whether an earlier age of the first graft elicits better vision results or reduces the survival rate of grafts is also not known.
We enrolled 29 infants with PA (which is the most common indication for PK in Chinese infants with CCOs). All infants received PK before 1 year of age. After surgery, optical correction and treatment for occlusion amblyopia were enforced strictly for all patients. We evaluated graft survival and visual outcome and identified the prognostic factors of these infants. The results of this study will help surgeons choose the suitable timing and methods of PK and predict the survival rates and vision outcomes of infants with PA.
MATERIALS AND METHODS
The study protocol followed the tenets of the Declaration of Helsinki (1975) and its later amendments. The study protocol was approved by the Ethics Review Board of Beijing Capital Medical University. Twenty-nine infants (37 eyes) with PA who underwent primary PK at Beijing Children's Hospital and Beijing Tongren Eye Center between April 2017 and March 2018 were reviewed.
Twenty-nine eyes underwent PK alone, 5 eyes underwent PK combined with lensectomy, and 3 eyes underwent lensectomy or vitrectomy several months after PK. The mean age at the time of PK was 5.7 ± 2.3 (range, 3–11) months. The mean duration of follow-up was 18 ± 3 (range, 11–22) months. Medical records, surgical information, and imaging data were collected.
The corneal graft was scored according to the criteria established by Dickey et al13: 0, graft is totally transparent; 1, slight haziness; 2, increased haziness, but anterior chamber structures are clearly visible; 3, advanced opacification with anterior chamber structures seen with difficulty; 4, opaque cornea, without the view of the anterior chamber. A corneal score of 2 to 4 was defined as graft failure. The duration of graft survival (determined from medical records) was from the time of surgery to the time of graft failure. All infants were examined by B-scan ultrasound and ultrasound biomicroscopy before surgery.
Donor corneas were obtained from Beijing Tongren Eye Bank. Donor corneal tissues with a healthy epithelium and high density of endothelial cells (≥2500 cells/mm2) were used. PK was performed under general anesthesia for all patients.
The corneal button was oversized from 0.75 mm to 1.0 mm, and the mean size was 7.85 ± 0.47 (range, 6.25–8.75 mm) depending on the lesion in the host cornea. Twelve or sixteen 10-0 nylon interrupted sutures were used. Additional procedures (eg, synechiolysis and cataract extraction) were undertaken as required. The bandage contact lens was placed on the corneal surface and left for 10 to 15 days. It was removed during examination using a handheld slit lamp while sleeping. Wearing a bandage contact lens at an early stage after surgery can reduce the symptoms of foreign body sensation, tearing, and the complications of epithelial defects (especially for cases without normal limbus). Suture removal was completed in most cases 6 to 8 weeks after surgery. Patients with a unilateral opacity were treated by occlusion of the fellow eye 2 hours per day (1 hour in the morning and afternoon, respectively) 2 weeks after surgery.9 All infants received cycloplegic refraction and had spectacle correction 2 weeks after suture removal.
Postoperative Care and Follow-up
A typical regimen involved tobramycin and dexamethasone eye drops (Tobradex; Alcon, Fort Worth, TX) 4 times daily and tobramycin and dexamethasone eye ointment (Tobradex) every night. Rb-bFGF eye gel (Essex Bio-Pharmaceutical, Zhuhai, China) was used twice daily for the first 2 weeks after surgery. Its main ingredients were recombinant bovine basic fibroblast growth factor and carbomer. It was used to improve healing of the transplanted corneal epithelium. Next, treatment with an ophthalmic suspension of 1% prednisolone acetate (Pred Forte; Allergan, Irvine, CA) 4 times daily with a gradual reduction in the following 12 months and a low-dose corticosteroid [fluorometholone (0.1%) eye drops; Santen, Tokyo, Japan] was maintained from 12 to 24 months after transplantation. Tobramycin (0.3%) eye drops (Tobrex; Alcon) were used as anti-infection treatment from 2 weeks after surgery to the time of suture removal. Tacrolimus (0.1%) eye drops (Talymus; Senju Pharmaceuticals, Osaka, Japan) were applied every 4 hours from postoperative day 3, with a gradual reduction in the following 24 months after transplantation.
In general, posttransplant examination was performed at 1 day, 1 week, and 1 month and then every 3 months, after surgery for 2 years. A handheld slit lamp, rebound tonometer (ICARE, Helsinki, Finland) and B-scan ultrasound were used at each follow-up. A hypnotic drug (chloral hydrate) was used if necessary.
All infants underwent assessment of visual acuity after surgery using Teller Acuity Cards II (Stereo Optical, Chicago, IL) when they were older than 6 months. Simultaneously, a red ball (diameter, 5 cm) was used to observe the follow-up distance. We observed children to assess if they could walk by themselves after 1 year of age. The results and criteria for visual acuity are shown in Supplemental Digital Content 1 (see Supplemental Table 1, http://links.lww.com/ICO/B166). All eyes were corrected with spectacles.
Statistical analyses were undertaken using SPSS (version 23; IBM, Armonk, NY). Descriptive statistics are reported. Survival of primary grafts was analyzed using the Kaplan–Meier method. Log-rank tests were used to evaluate differences in survival between groups.3 The prognostic factors for graft survival and visual outcome (surgical indication; unilateral versus bilateral grafts; age at which primary surgery was performed; and other ocular surgery methods) were tested using the χ2 test. P < 0.05 was considered significant.
Thirty-seven eyes of 29 infants (11 boys, 18 girls) underwent primary PK during the study period. The mean age at the time of PK was 5.7 ± 2.3 (range, 3–7) months. All infants received PK once during follow-up. Eight of 12 patients with bilateral opacities had bilateral corneal transplantation. The mean duration of follow-up was 18 ± 3 (range, 11–22) months. The mean age at the last follow-up visit was 23 ± 4 (range, 15–30) months.
We divided 37 eyes into 3 types according to the corneal lesion: nonvascularized PA I (nv-PA I), vascularized PA I (v-PA I), and PA II (Fig. 1). The classification was based on the Nischal14 approach for neonatal corneal opacities.
Until the final follow-up of 37 eyes, grafts in 27 eyes (73%) maintained clarity (score of ≤1). Kaplan–Meier survival analyses indicated the prevalence of overall graft survival at 1 year to be 78.4%.
At the final follow-up, the grafts of nv-PA I maintained the highest prevalence of transparency (92.9%; 13/14), whereas that of v-PA I was 75% (9/12 eyes) and that of PA II was 45.5% (5/11) (Table 1). There was a significant difference among the 3 groups (χ2 = 6.391, P = 0.041). Kaplan–Meier survival curves indicated the prevalence of graft survival at 1 year to be 92.9%, 75%, and 63.6% for nv-PA I, v-PA I, and PA II, respectively (Fig. 2).
TABLE 1. -
Analysis of Graft Survival and Visual Outcome for Infant PK
||No. of Eyes
||Clear Grafts, %
||Visual Acuity, % (n)
| Nonvascularized Peters I
| Vascularized Peters I
| Peters II
|Laterality of surgery
| Unilateral PK
| Bilateral PKs
|Age at PK
| Younger than 6 mo
| 6 mo or older
*Including 4 eyes with bilateral opacities that received only unilateral PK.
Seventeen infants with a unilateral opacity underwent unilateral PK. Eight of the 12 patients with bilateral opacities underwent bilateral PKs, and another 4 patients with bilateral opacities received only unilateral PK because of very light opacity (3 eyes) or original retinal detachment (1 eye). The prevalence of graft survival for infants who had unilateral PK was 76.2% (16/21), and for infants who had bilateral PKs, it was 68.8%; and this difference was not significant (χ2 = 1.763, P = 0.184). Of the 16 eyes of 8 infants who underwent bilateral PKs, 5 grafts experienced reopacity. Three of them were due to rejection, and the other 2 grafts failed because of an epithelial defect secondary to opacities (Table 1).
According to age at the time of primary PK, infants were divided into 2 groups: younger than 6 months (21 eyes) and 6 months or older (16 eyes). We found that 76.2% (16/21) and 68.8% (11/16) grafts remained transparent, but an association with graft survival was not found (χ2 = 0.199, P = 0.656) (Table 1).
We wished to investigate the influence of the surgical procedure on the prevalence of graft survival. Hence, we compared the prevalence of graft survival between patients who underwent PK only and those who received other types of surgical procedure. The prevalence of graft survival in eyes that underwent PK only was 79.3% (23/29). Five eyes that underwent PK combined with extracapsular cataract extraction (ECCE) simultaneously demonstrated a low prevalence of graft survival (40%; 2/5). Two eyes that underwent ECCE 16 months after the first PK had grafts that remained transparent. The graft failed in 1 eye that underwent ECCE and vitrectomy 4 months after the first PK.
Of the 37 enrolled eyes, 25 (67.6%) eyes achieved ambulatory vision or better. Eighteen (48.7%) of those eyes had postoperative visual acuity ≥20/260. Twelve (32.4%) eyes were defined as having nonambulatory vision (including 6 eyes of 6 patients with a unilateral opacity who refused to cover their fellow eye and could not obtain monocular vision). For all clear grafts, 85.3% (23/27) of eyes achieved ambulatory vision or better.
A lower proportion of the v-PA I group (66.7%) and PA II group (36.4%) achieved ambulatory vision than the nv-PA I group (78.6%) (P = 0.041). For visual acuity >20/260, nv-PA I, v-PA I, and PA II showed the same trend, reaching values of 71.4%, 50.0%, and 18.2%, respectively (P = 0.030) (Table 1).
Among all 16 eyes with bilateral opacities that received bilateral PK, 12 eyes (75.0%) achieved ambulatory vision. Among the other 21 eyes (17 with unilateral opacity and 4 with bilateral opacities) that received only unilateral PK, 13 eyes (61.9%) achieved ambulatory vision. There were no significant differences between them. However, regarding visual acuity >20/260, infants with bilateral opacities who had bilateral PKs showed better outcomes (68.8%) than those with unilateral PK (33.3%) (P = 0.049) (Table 1). There were no significant differences for visual results in the 2 age groups (younger than 6 months and 6 months or older; Table 1).
We also compared visual acuity between infants who underwent PK only and those who had PK combined with other types of surgery. In all 29 infants who underwent PK only, 75.9% of eyes achieved ambulatory vision and 55.2% of eyes achieved visual acuity of 20/260. Five eyes underwent PK combined with ECCE, and only 1 eye of those eyes achieved ambulatory vision. Two eyes were found to have complicated cataract 1 year after PK; they underwent ECCE and achieved visual acuity of 20/260. One eye that underwent ECCE and vitrectomy did not achieve ambulatory vision.
Graft Failure and Complications
Ten grafts (27%) failed. Graft rejection occurred in 8 eyes, and a continuous corneal epithelial defect secondary to opacities occurred in another 2 eyes (see Discussion). Two grafts of the 8 rejected grafts recovered from a score of 4 to 2 after medical treatment and, at the end of follow-up, they achieved visual acuity of 20/260 (see Supplemental Table 2, Supplemental Digital Content 2, http://links.lww.com/ICO/B167).
Seven eyes (18.9%) experienced glaucoma 4 to 19 (mean, 11) months after surgery, and until the final follow-up, all cases were controlled by antiglaucoma medication (timolol eye drops, brinzolamide eye drops, and latanoprost eye drops). Single or combined use of antiglaucoma drugs mainly depended on the intraocular pressure (IOP) and the growth rate of axial length. At the last follow-up, only 2 eyes (28.6%) achieved 20/260, 1 eye (14.3%) achieved 20/470, and the other 4 eyes (57.1%) did not achieve ambulatory vision.
Five eyes (13.5%) had a complicated cataract. One was a topical posterior capsule opacification that occurred 14 months postoperatively and did not undergo additional surgical treatment. Three eyes had total cataracts; 2 of them with clear grafts received ECCE 16 months after PK, and the other eye had graft rejection and did not receive further surgery. The final eye with a complicated cataract had graft rejection and underwent ECCE and vitrectomy for adhesion of the lens capsule with the peripheral retina.
The period 0 to 3 years after birth is critical for visual development.11 The sensory stereopsis is established in most infants at 20 to 28 weeks.15 Visual deprivation during the early months of life because of CCOs can result in long-term changes to the central nervous system. This may result in profound and uncorrectable loss of vision that can impact negatively on a child's development.6 Evidence shows that deprivation amblyopia has an effect on more than just the visual cortex: there is also an effect on the dorsal pathway, including the cerebellum.16,17
Early PK is the first-line treatment to save the visual function of children.8 In 1999, Yang et al4 reviewed 47 patients with PA (72 eyes): the mean age at the first PK was 4.4 months, and graft survival at 1 year was 49%. In 2001, Richard et al6 reported 11 children with a CCO who received PK before the age of 13 months; graft survival was 61% at 1 year after the first transplant. In 2011, Basdekidou et al9 reviewed 14 patients with unilateral PA who underwent PK at an average age of 9 months. At the end of 30-month follow-up, 78.6% grafts remained transparent. The surgical technology of PK for infants with CCO is progressing constantly.18,19
Corneal transplantation for infants started relatively late in China, and most surgeons choose to undertake corneal transplantation at the age of 1 year or older. In this study, 29 infants with PA aged 3 to 11 months underwent PK; 73% of the eyes remained transparent at the end of 18-month follow-up, better than our previous reports in children.7 This finding may be related to continuous improvement of surgical methods and long-term use of the new, local immunosuppressants. All infants began to receive topical calcineurin inhibitors (0.1% tacrolimus eye drops) from postoperative day 3 to at least 2 years after PK. Tacrolimus is a potent immunosuppressive agent used to prevent tissue rejection, and obvious topical or systemic complications have not been associated with its use.20,21
Of 10 failed grafts, 8 grafts showed rejection. After 4 to 10 months of drug treatment22 for graft rejection, 2 grafts recovered partial transparency in the pupil area, and ambulatory vision was restored. For graft rejection in infants, early treatment is very important, and treatment should last 1 to 3 months. The next transplantation should be considered carefully.23 The other 2 cases of graft failure had a persistent epithelial defect and were due to a lack of normal limbus before surgery. It is very difficult to elicit healing by drug treatment and a bandage contact lens compared with overcoming a secondary opacity and graft failure. Our experience is that an amniotic membrane graft is not very effective for such complications.
We oversized grafts by 0.75 to 1 mm to avoid the possibility of a postoperative shallow anterior chamber and anterior synechiae. The corneoscleral tissue of infants was very thin, so oversizing the grafts could avoid compressing the recipient tissue. In addition, the curvature of an infant's cornea is significantly higher than that of an adult. A postoperative curvature of this size is suitable for an infant's physiological state, and our study of the postoperative refractive index did not show obviously high myopia.
High intraocular pressure (IOP) is a major complication that concerns surgeons after corneal transplantation in infants and young children.5 However, IOP measurement in infants is also affected by graft thickness, and it cannot be used as the only indicator of glaucoma. Because of the high elasticity of sclera, infants often show eyeball dilation or corneoscleral staphyloma when faced with a high IOP. The length of the eye axis increases rapidly, and exophthalmos often appears before optic nerve damage and becomes a warning of a high IOP in infants. Antiglaucoma eye drops should be used if an infant shows a high IOP and an abnormal increase in the eye axis length. If these drugs cannot control it, infants will receive cyclocryotherapy or trabeculectomy.
For infants, a sustained high IOP will damage corneal endothelial cells, resulting in opacity and graft edema. Vision will decrease due to graft edema, which is more obvious than the damage to the optic nerve due to glaucoma.
A complicated cataract should be treated according to its severity. Five eyes that underwent PK combined with ECCE did not have visual acuity of >20/260. For children (especially infants who mainly use near vision), lens removal is not recommended if the cataract does not completely block the visual axis.
We referred to Nischal classification of CCOs14,16,24 combined with intraoperative findings and ultrasound biomicroscopy to divide all cases into 3 groups: nv-PA I, v-PA I, and PA II. The prevalence of graft transparency was 92.9%, 75%, and 45.5%, respectively. Corneal vascularization and abnormal lens were the main differences between the 3 groups. Preoperative corneal stromal vessels, as risk factors for corneal transplant failure, have been reported in CCOs and acquired corneal opacities.1,19,25 Similarly, there was a significant difference in the proportion of eyes with ambulatory vision and visual acuity >20/260 in the 3 groups. The PA II group had the worst vision results, even though they were treated early with optical correction and occlusion. The nv-PA I group with a normal lens and few corneal vessels showed the best results.
Early visual development after birth has a high degree of plasticity. The theory that deprivation amblyopia is irreversible in infants has been shown to be incorrect.9,17 Basdekidou et al9 reported that 14 patients who underwent keratoplasty for unilateral PA achieved useful vision results. Early surgery to open the visual pathway, refractive correction, and strict amblyopia training are critical for the vision development of infants. We believe that, for the 6 children who refused to cover their eyes for visual acuity assessment, their eyes had poor vision and faced a strong irritation phenomenon. But this does not mean that they did not obtain any vision. If the grafts remain transparent and their parents adhere to amblyopia treatment at the early stage, they have chance to obtain some vision.
Vision outcome after PK for bilateral opacities was better than that for a unilateral opacity. We hypothesize that this result was because deprivation of bilateral vision at birth keeps the visual cortex in an immature state, and the plasticity period of visual cortex development is extended.26 Moreover, unilateral opacity must face binocular competition. Hence, PKs for bilateral opacities will result in greater vision improvement in children. Even though PKs for infants are more likely to result in graft failure, visual improvement will promote global development in behavior, communication, and ambulation, even if the graft fails at 12 to 18 months.27,28
According to the age at which primary PK was performed, patients were divided into 2 groups: younger than 6 months and 6 months or older. There was no significant difference in the prevalence of graft survival and vision outcome between these 2 groups. These data are different to results from a study by Kavita et al.8 They found that poor graft survival correlated with age younger than 6 months. They postulated this was due to the difficulty in examining and late recognition of complications. In this study, strict follow-up and detailed examination did not show a significant difference in the prevalence of graft rejection in the 2 age groups. Another reason may be that the age at which surgery was performed in the study by Kavita et al was 2.5 to 22 months, whereas it was 3 to 11 months in this study.
In summary, for infants with PA who underwent PK, the prevalence of graft survival and visual acuity were related mainly to the indication. The main risk factors were corneal vascularization and an abnormal lens. Early surgery and occlusion treatment may prevent deprivation amblyopia and achieve better visual outcomes.
The authors thank the patients and their families for taking part in this study.
1. Lowe MT, Keane MC, Coster DJ, et al. The outcome of corneal transplantation in infants, children, and adolescents. Ophthalmology. 2011;118:492–497.
2. Ganekal S, Gangangouda C, Dorairaj S, et al. Early outcomes of primary pediatric keratoplasty in patients with acquired, atraumatic corneal pathology. J AAPOS. 2011;15:353–355.
3. Karadag R, Chan TC, Azari AA, et al. Survival of primary penetrating keratoplasty in children. Am J Ophthalmol. 2016;171:95–100.
4. 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.
5. Huang C, O'Hara M, Mannis MJ. Primary pediatric keratoplasty: indications and outcomes. Cornea. 2009;28:1003–1008.
6. Comer RM, Daya SM, O'Keefe M. Penetrating keratoplasty in infants. J AAPOS. 2001;5:285–290.
7. Patel HY, Ormonde S, Brookes NH, et al. The indications and outcome of paediatric corneal transplantation in New Zealand: 1991-2003. Br J Ophthalmol. 2005;89:404–408.
8. Rao KV, Fernandes M, Gangopadhyay N, et al. Outcome of penetrating keratoplasty for Peters anomaly. Cornea. 2008;27:749–753.
9. Basdekidou C, Dureau P, Edelson C, et al. Should unilateral congenital corneal opacities in Peters' anomaly be grafted? Eur J Ophthalmol. 2011;21:695–699.
10. Kim YW, Choi HJ, Kim MK, et al. Clinical outcome of penetrating keratoplasty in patients 5 years or younger: Peters anomaly versus sclerocornea. Cornea. 2013;32:1432–1436.
11. Kiorpes L, Kiper DC, O'Keefe LP, et al. Neuronal correlates of amblyopia in the visual cortex of macaque monkeys with experimental strabismus and anisometropia. J Neurosci. 1998;18:6411–6424.
12. Lin Q, Shi W, Miao S, et al. Visual outcomes and prognostic factors of successful penetrating keratoplasty in 0- to 7-year-old children with congenital corneal opacities. Cornea. 2018;37:1237–1242.
13. Dickey JB, Cassidy EM, Bouchard CS. Periocular FK-506 delays allograft rejection in rat penetrating keratoplasty. Cornea. 1993;12:204–207.
14. Nischal KK. A new approach to the classification of neonatal corneal opacities. Curr Opin Ophthalmol. 2012;23:344–354.
15. Elston JS, Timms C. Clinical evidence for the onset of the sensitive period in infancy. Br J Ophthalmol. 1992;76:327–328.
16. Mannis MJ, Holland EJ. Cornea. 4th ed. Edinburgh, NY: Elsevier; 2017.
17. Ostrovsky Y, Andalman A, Sinha P. Vision following extended congenital blindness. Psychol Sci. 2006;17:1009–1014.
18. Medsinge A, Speedwell L, Nischal KK. Defining success in infant penetrating keratoplasty for developmental corneal opacities. Am Orthopt J. 2014;64:81–88.
19. Majander A, Kivela TT, Krootila K. Indications and outcomes of keratoplasties in children during a 40-year period. Acta Ophthalmol. 2016;94:618–624.
20. Magalhaes OA, Marinho DR, Kwitko S, et al. Topical 0.03% tacrolimus preventing rejection in high-risk corneal transplantation: a cohort study. Br J Ophthalmol. 2013;97:1395–1398.
21. Hashemian MN, Latifi G, Ghaffari R, et al. Topical tacrolimus as adjuvant therapy to corticosteroids in acute endothelial graft rejection after penetrating keratoplasty: a randomized controlled trial. Cornea. 2018;37:307–312.
22. 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.
23. Holland EJ, Chan CC, Wetzig RP, et al. Clinical and immunohistologic studies of corneal rejection in the rat penetrating keratoplasty model. Cornea. 1991;10:374–380.
24. Nischal KK. Genetics of congenital corneal opacification—impact on diagnosis and treatment. Cornea. 2015;34(suppl 10):S24–S34.
25. Low JR, Anshu A, Tan AC, et al. The outcomes of primary pediatric keratoplasty in Singapore. Am J Ophthalmol. 2014;158:496–502.
26. ZQ Y, Tao Y. Update on binocular form deprivation amblyopia: mechanisms and treatment. Chin J Optom Ophthalmol Vis Sci. 2010;12:321–323.
27. Dale NJ, Tadic V, Sonksen P. Social communicative variation in 1-3-year-olds with severe visual impairment. Child Care Health Dev. 2014;40:158–164.
28. Stulting RD, Sumers KD, Cavanagh HD, et al. Penetrating keratoplasty in children. Ophthalmology. 1984;91:1222–1230.