The Effect of Topical Autologous Serum on Graft Re-epithelialization After Penetrating Keratoplasty


To analyze factors influencing corneal graft re-epithelialization after penetrating keratoplasty (PK) and evaluate the effect of topical autologous serum in promoting graft re-epithelialization.


Prospective interventional study.


We analyzed 165 eyes of 165 patients who underwent PK between January 1, 2005 and December 31, 2007. Patients were divided into 2 groups according to routine use or non-use of postoperative 20% topical autologous serum. Postoperative slit-lamp examination after fluorescein staining was performed, and graft re-epithelialization time was recorded. Recipient/donor characteristics, surgical variables, and topical use of autologous serum were analyzed for their effects on post-PK graft re-epithelialization. Statistical analysis was performed by univariate and multivariate regression analysis using the ordinal logistic fit model to assess the potential risk factors influencing graft re-epithelialization after PK.


In univariate analysis, diabetes mellitus (DM), longer death-to-storage time and death-to-surgery time of the donor, and larger recipient size significantly delayed graft re-epithelialization ( P < .05). Use of autologous serum significantly expedited graft re-epithelialization ( P = .004). In multiple regression analysis, only DM in the recipient (odds ratio [OR] = 5.10, P < .001), postoperative use of autologous serum (OR = 0.54, P = .046), and larger graft size (OR = 4.44, P < .001) influenced graft re-epithelialization. The beneficial and healing effect of autologous serum is particularly significant in diabetic recipients and larger grafts.


Several factors may influence graft re-epithelialization after PK. Graft re-epithelialization time was longer in diabetic recipients and larger grafts. Use of autologous serum may be a beneficial strategy in these patients with potentially delayed epithelial healing.

During the postoperative course of penetrating keratoplasty (PK), rapid graft re-epithelialization and maintenance of an intact epithelium are critical for visual acuity, graft transparency, graft survival, and protection of the stroma against infection and melting. Graft re-epithelialization usually takes no longer than 1 week after transplantation. Chronic epithelial defects lead to inflammation, infection, corneal scarring, corneal melting, perforation, and even graft failure. Several factors associated with graft re-epithelialization after PK using organ-cultured donor tissue stored at 31°C to 37°C have been reported. However, similar studies of hypothermically preserved corneal buttons are limited. A better understanding of these factors will help us prevent postoperative epithelial defects and complications.

Topical autologous serum had been used clinically as an in vivo topical medication in treating ocular surface disorders such as recurrent corneal erosion, persistent epithelial defects, superior limbic keratoconjunctivitis, and dry eyes. The benefits of using autologous serum to promote corneal epithelial healing or for treatment of persistent corneal epithelial defects have been described. Persistent epithelial defects can be caused by a variety of insults, including keratoconjunctivitis sicca (KCS), exposure keratopathy, neurotrophic ulcer, corneal stem cell failure, and postinfectious ulcers. Serum, the clear liquid part of full blood that remains after cellular components and clotting proteins have been removed, is thought to be better than artificial tear substitutes for the following reasons. First, it contains essential ocular surface nutrients—cytokines, proteins, growth factors, lipids, vitamins, and bacteriostatic components—that are lacking in artificial tears but exist in normal tear film. Second, its pH, osmolarity, and biomechanical properties are similar to natural tears. Third, it has no preservatives, stabilizers, or other additives that are often present in artificial tears and may increase corneal epithelial toxicity. Use of autologous serum has been advocated in patients requiring lamellar keratoplasty or diabetic vitrectomy in whom problems with epithelial healing are anticipated. Similar studies on autologous serum and umbilical serum therapy for persistent corneal epithelial defects in patients after PK have also been reported. However, there have been limited large prospective studies regarding the effect of autologous serum on graft re-epithelialization after PK. Therefore, we conducted a prospective study to analyze factors influencing corneal graft re-epithelialization after PK by using hypothermically preserved corneal buttons and to evaluate the possible beneficial effects of autologous serum in promoting graft re-epithelialization in these patients.


Study Design

We conducted a prospective interventional study with 174 patients who underwent PK performed by 2 surgeons (W.L.C. and F.R.H.) at National Taiwan University Hospital (NTUH) between January 1, 2005 and December 31, 2007. The criteria for enrollment were: patients who received PK with no ocular or systemic disease that may affect epithelial healing (eg, severe dry eye, severe lid abnormalities, limbal stem cell deficiency, or corneal anesthesia) except diabetes mellitus (DM). Nine of the 174 patients (5%) received 2 or more grafts in the same or contralateral eye during the study period. We analyzed only the first graft performed during the study period. A total of 165 PK procedures performed in 165 patients were included in this study.

Donor Corneas

Donor corneas were either obtained from domestic donors or imported from the United States. Among the imported corneal buttons, most (83.3% [90/108]) were provided by the Cincinnati Eye Bank (Cincinnati, Ohio, USA). All of the corneal buttons were preserved in a viewing chamber filled with Optisol GS (Chiron Intraoptics, Irvine, California, USA) at 4°C using standard eye bank procedures before the operations. The donor corneal buttons were inspected under slit-lamp biomicroscopy for evaluation of integrity, clarity, and transparency before being qualified for use in the PK surgeries.


Patient characteristics are shown in Table 1 . DM was defined in this study as 1) nonfasting glucose levels of ≥200 mg/dL (11.1 mmol/L) or 2) a physician’s diagnosis of diabetes and use of diabetic medications.


Corneal Re-epithelialization After Penetrating Keratoplasty: The Characteristics of Recipients Classified by the Presence of Diabetes Mellitus

Diabetic Recipients Nondiabetic Recipients
Recipient Characteristics All Recipients With Autologous Serum No Autologous Serum With Autologous Serum No Autologous Serum
Patient numbers 165 34 42 48 41
Gender (male/female) 79/86 15/19 20/22 23/25 21/20
Age (years) 60.3 ± 19.3 60.4 ± 15.1 58.8 ± 14.6 61.0 ± 20.3 60.9 ± 19.5
Range 10–87 29–80 32–87 10–82 17–81
Preoperative diagnosis Patient numbers (%)
Bullous keratopathy 67 (40.6) 15 20 19 13
Corneal scar, corneal ulcer, interstitial keratitis 46 (27.9) 12 11 13 10
Keratoconus 15 (9.1) 1 2 6 6
Fuchs dystrophy 9 (5.5) 1 3 2 3
Other corneal dystrophy 4 (2.4) 1 0 1 2
Regraft 17 (10.3) 3 5 4 5
Miscellaneous preoperative diagnosis 7 (4.2) 1 1 3 2
Postoperative lens status
Phakic 69 (41.8) 9 12 25 23
Posterior chamber intraocular lens 79 (47.9) 22 26 16 15
Anterior chamber intraocular lens 13 (7.9) 2 3 5 3
Aphakic 4 (2.4) 1 1 2 0
Diabetes mellitus 76 (46.1)
Preoperative glaucoma history/pre-existing glaucoma 21 (12.7) 4 5 7 5
Previous glaucoma surgery 6 (3.6) 1 2 2 1
Preoperative use of antiglaucoma drops 18 (10.9) 3 5 6 4
Postoperative intraocular pressure >20 mm Hg 32 (19.4) 6 9 9 8
Postoperative use of antiglaucoma drops 24 (14.5) 5 7 7 5
Combined procedure: cataract surgery 47 (28.5) 11 12 13 11
Postoperative use of autologous serum 82 (49.7)

Surgical Procedures

All operations were performed at a single institution (NTUH) by 2 surgeons (W.L.C. and F.R.H.) using standard surgical procedures reported elsewhere. Briefly, donor corneas were punched out from the posterior corneal surface using a disposable donor cornea punch (Moria, Antony, France). Recipient corneas were trephinated using a disposable Hanna trephine (Moria), except in some cases of regrafts. The trephinated size of the donor cornea was 0.5 mm larger than that of the recipient, except in cases of keratoconus, in which the donor cornea was either the same as or 0.25 mm larger than that of the recipient. The graft was secured with a 10-0 nylon suture with either 16 interrupted sutures or 8 interrupted sutures combined with a single-running suture. The corneal epithelium was removed completely with cellulose sponges routinely after each operation. No bandage contact lens was inserted at the end of the operation. After surgery, all patients received 0.1% betamethasone, 0.3% gentamicin, and 0.4% tropicamide (Mydrin M; Santen, Osaka, Japan) application 4 times daily and dexamethasone/neomycin/polymyxin B sulfate ointment (Maxitrol; Alcon, Puurs, Belgium) once daily at night.

Additional Postoperative Treatment

Patients were divided into the autologous serum treatment group or the nonautologous serum treatment group according to surgical time. From January 1, 2005, to June 30, 2006, all 83 eyes of 83 patients (50.3% of the 165 total eyes in this study) were assigned to the nonautologous serum treatment group. From July 1, 2006, to December 31, 2007, all 82 eyes of 82 patients (49.7% of total eyes) were assigned to the autologous serum usage group, and topical usage of 20% autologous serum eye drops bi-hourly during waking hours in the lesion eye was prescribed. The autologous serum was prepared according to the procedure described by Tsubota and associates, except for the use of preservative-free artificial tear eye drops instead of saline for autologous serum dilution in our study. In this study, we discontinued the autologous serum usage once the complete graft re-epithelialization was achieved.

Main Outcome Measure

Corneal epithelial healing was recorded daily by slit-lamp examination with fluorescein staining. Patients were hospitalized and examined daily for graft re-epithelialization, which was the main outcome measure. Patients with postoperative chronic persistent epithelial defects for >14 days after the operation were treated with therapeutic contact lens (TCL) application and followed up as outpatients. Graft re-epithelialization time was recorded and classified into 5 grades: grade I, complete graft re-epithelialization within 3 days of the operation; grade II, graft re-epithelialization within 4 to 6 days; grade III, graft re-epithelialization within 7 to 10 days; grade IV, graft re-epithelialization within 11 to 14 days; and grade V, persistent epithelial defects for >14 days after operation, in which case a TCL was applied.

Statistical Analysis

All data were analyzed via a commercially available statistical software package (Stata version 8.2; Stata Corp, College Station, Texas, USA). Related factors for graft re-epithelialization that were first examined via univariate analysis included the following: donor characteristics (age and gender, cause of death, death-to-storage time, death-to-surgery time, and origin [either from fresh donors in Taiwan or from eye banks in the United States]; recipient characteristics (age and gender, surgical indications, diabetic history, pre-existing glaucoma, previous glaucoma surgery, and antiglaucoma eye drop use); surgical variables (donor trephination size, suture method [interrupted or mixed sutures], and combined cataract surgeries; postoperative variables (postoperative high intraocular pressure [>20 mm Hg], postoperative antiglaucoma eye drop use, and postoperative use of autologous serum; and other factors, including hyphema, wound leakage, or vitreoretinal complications. Variables with P < .05 in univariate analysis were selected for multivariate analysis. Univariate and multivariate regression analyses were performed using the ordinal logistic fit model to assess the effects of these variables on postoperative graft re-epithelialization.


Of the 165 eyes, 108 corneal grafts (65.5%) were of international origin and 57 (34.5%) were of domestic origin. The average donor age was 67.8 ± 11.2 years (mean ± standard deviation), average death-to-storage time was 15.5 ± 5.6 hours (range, 4-26 hours; 10.6 ± 4.0 hours for domestic corneas and 18.1 ± 4.4 hours for imported corneas), and average death-to-surgery time was 5.9 ± 3.3 days (range, 1–13 days; 2.3 ± 0.9 days for domestic corneas and 7.8 ± 2.4 days for imported corneas). Donor causes of death were cancer (33.3% [55/105]), cardiovascular disease (22.4% [37/165]), intracerebral hemorrhage (18.8% [31/165]), respiratory disease (6.1% [10/165]), and miscellaneous disease (19.4% [32/165]).

History of DM was found in 46.1% (76/165) of patients. One hundred thirty-five eyes received interrupted sutures while 30 eyes received mixed interrupted and continuous sutures during graft suturing. The average graft size was 7.8 ± 0.5 mm (range, 7.0-9.0 mm). The characteristics of all recipients classified by the presence of DM are listed in Table 1 .

Complete corneal epithelial healing was achieved within 14 days in 151 of 165 eyes (91.5%). Fourteen of 165 eyes (8.5%) eyes had postoperative epithelial defects that persisted for more than 14 days and were treated with TCL. All corneal epithelia healed within 3 to 16 weeks after surgery in this group. In univariate analysis, death-to-storage time, death-to-surgery time, trephination size, use of autologous serum, and DM of the recipient all significantly influenced graft re-epithelialization time ( P < .05, Table 2 ). In multiple regression analysis, only trephination size (odds ratio [OR] = 4.44, P < .001) and DM of the recipient (OR = 5.10, P < .001) significantly prolonged graft re-epithelialization ( Table 2 ). The graft epithelial healing time was shorter in patients who received postoperative autologous serum applications (OR = 0.54, P = .046). Among the 82 patients who received the postoperative autologous serum, only 4 (4.9%) had a persistent epithelial defect 14 days after surgery that required TCL use before complete corneal re-epithelialization, while 10 out of the 83 patients (12.0%) not using the autologous serum required TCL use postoperatively. In addition, we also found that diabetic patients tended to have postoperative persistent epithelial defects requiring TCL use (12 of 14 patients, 85.7%). In diabetic recipients using autologous serum, 58.8% (20/34) of eyes achieved complete corneal re-epithelialization in 7 days and 82.4% (28/24) in 10 days. In diabetic recipients not using autologous serum, 33.3% (14/42) of eyes achieved complete corneal re-epithelialization in 7 days and 45.2% (19/42) in 10 days. The beneficial effect of autologous serum in promoting epithelial healing is particularly significant in diabetic recipients ( Figure 1 ). Autologus serum is also found to promote epithelial healing in larger grafts ( Figure 2 ). Representative pictures of postoperative corneal re-epithelialization in 2 diabetic recipients are shown in Figure 3 .

Jan 17, 2017 | Posted by in OPHTHALMOLOGY | Comments Off on The Effect of Topical Autologous Serum on Graft Re-epithelialization After Penetrating Keratoplasty

Full access? Get Clinical Tree

Get Clinical Tree app for offline access