The present retrospective study included 20 patients who underwent TPK surgery between January and May in 2017 in the Beijing Tongren Eye center (Beijing, China). The patients comprised 14 males and 6 females. The average age was 57.1±10.77 years (range, 40-70 years). Patients included in the current study underwent TPK and had positive corneal scraping culture results. Patients suspected of bacterial keratitis routinely underwent aerobic culture of the specimen. Corneal scrapings were inoculated onto nutrient broth, blood agar plate (0.05% Defibrinated Sheep Blood), MacConkey and Chocolate Agar Plate (0.05% Defibrinated Sheep Blood) medium (Beijing Solarbio Science & Technology Co., Ltd.). Subsequently, samples were incubated for 1-2 days at 28˚C and 50% humidity and isolated individual colonies. The Kirby-Bauer method (Disc diffusion method) was used to determine antibiotic sensitivity of different antibiotics. The antibiotics tested were all purchased from Beijing Solarbio Science & Technology Co., Ltd. and included Amikacin (30 µg), Ceftazidime (30 µg), Cefoxitin (30 µg), Ciprofloxacin (5 µg), Gentamicin (120 µg), Ofloxacin (5 µg), Levofloxacin (5 µg), Rifampicin (5 µg), Moxifloxacin (5 µg), Gatifloxacin (5 µg), Oxacillin (1 µg). Potato Dextrose Agar (PDA) (Beijing Luqiao Technology Co., Ltd.) was selected as medium for specimens suspected as fungal keratits. Specimens were placed in Thermostatic-Humidistat Cultivating Box at 28˚C and 40% humidity for 3 to 10 days. After that, the colonies in the positive specimens were isolated and transferred to PDA dishes (Tianjin Jinzhang Company). The Rosco paper diffusion method (Neo-Sensitab drug sensitive paper; Rosco Diagnostica) was used for drug sensitivity test. Test drugs include: Natamycin (50 µg; Rosco Diagnostica), Terbinafine (30 µg; Rosco Diagnostica), Itraconazole (8 µg; Rosco Diagnostica), Fluconazole (15 µg; Rosco Diagnostica), Amphotericin B (15 µg, Rosco Diagnostica), Voriconazole (1 µg; Rosco Diagnostica). The diagnostic criteria of severe infectious keratitis (1) was as follows: Corneal ulcer area >6 mm² or lesion depth >2/3 of corneal thickness; presence of anterior chamber exudate or empyema; and corneal ulcer perforation or descemetocele. The criteria for surgery for patients with severe infectious keratitis (15) were: No healing after 7 days of standard treatment; descematocele with infiltrates; and corneal perforation >3 mm with active inflammation. All corneal transplants were performed by the same surgeon (JY) in Beijing Tongren Eye center (Beijing, China). Excluded patients had a history of previous corneal transplant and/or had congenital eye pathologies, such as cornea dystrophy or leucoma.

The patients were diagnosed with bacterial keratitis (n=8) or fungal keratitis (n=12). Each diagnosis was supported by a positive culture in all cases. Pre-operative assessment indicated that the patients exhibited iris adhesions (n=4), high intraocular pressure (n=7) and anterior chamber empyema (n=12). Additionally, 17 eyes presented with pre-operative corneal neovascularization and a single eye exhibited descemetocele. The mean recipient bed and recipient graft were 7.8±1.3 and 8.3±1.2, respectively. Pre-operative data are presented in Table I. Due to corneal ulcers problems, Digital Tonometry was used as intraocular pressure measurement. (Tn represented normal pressure. T+1, T+2 and T+3 indicated different degrees of intraocular pressure increase, with T+3 as the highest). Corneal perforation or descemetocele was denoted as ‘1 point’, and the absence of these was denoted as ‘0 points’. A lack of iris adhesion was recorded as ‘0 point’. If iris adhesion was it was assigned a score of 1 to 4 quadrants, which was correspondingly recorded as ‘1-4 points’. The height of hypopyon (h) was based on the vertical diameter of the cornea. Ulcer depth was based on central cornea thickness. Ulcer size, ulcer depth and the height of hypopyon were measured using Image Processing Software (vision 11.0; EdgeView, Apple Corporation). TPK sites were divided into orthotopic (O) and partial (P) transplantation. Corneal vascularization was divided into 1-4: 1, ≤25% corneal diameter; 2, ≤50% corneal diameter; 3, ≤75% corneal diameter; 4, ≤100% corneal diameter. For example, (2,1) represents corneal vascularization involving 2 quadrants, <25% of the corneal diameter.

All corneal transplants were performed with similar technique by the same surgeon at Tongren Eye Hospital (14,16). Donor corneas were stored in OptiSol storage media (Bausch & Lomb Co., Ltd.) for 7-10 days at 4˚C. The host trephine size was selected to completely cover the infiltrate edge of the ulcer and was oversized by 0.25 mm. Anterior chamber washout was performed in patients with anterior chamber hypopyon as previously described by Jia et al (17). Furthermore, patients with synechia or severe anterior chamber empyema prior to surgery underwent peripheral iridotomy at the 12 o'clock position at the time of surgery (18). The cornea grafts were sutured to the host with 16 interrupted 10-0 monofilament nylon sutures. Suture knots were trimmed and buried toward the donor side. In all cases, the excised host cornea was sent for microbiologic and histopathologic examination in Beijing Tongren Eye center (Beijing, China).

Selection of the appropriate antibiotic eye drops and topical antimicrobial agents was based on the microbiological profile and specific sensitivities of the causative infective agent (19).

Quinolone eye drops were administered to all patients as the culture results indicated that gram positive cocci infections were present. The current study administered 5 ml 0.5% gatifloxacin eye drops (Anhui Shuangke Pharmaceutical Co., Ltd.) 4 times/day with 5 g 0.3% gatifloxacin gel (Shenyang Xingqi Eye Medicine Co., Ltd.) 3 times/day for 2 weeks then gradually reduced the amount according to corneal condition.

All cases of fungal infection were treated with 5 ml 5% natamycin eye drops (Hubei Pharmaceutical Co., Ltd.) and 0.5% fluconazole eye drops (Yongguang Pharmaceutical Co., Ltd.) 4 times/day for 2 weeks then gradually reduced the amount according to corneal condition.

Furthermore, 5 mg 0.1% tacrolimus eye drops (Senju Pharmaceutical Co., Ltd.) were applied 4 times/day for 2 months then gradually reduced the amount according to corneal condition starting on the first day post-surgery for all patients.

Steroid eye drops were not administered during the first post-operative month. However, after a period of one month, if no signs of active inflammation were present, 1% prednisolone acetate (Allergan Pharmaceutical Co., Ltd.) was added to the treatment regimen and applied 4 times/day for a month then gradually reduced the amount according to corneal condition. Additionally, 0.3% tobramycin/0.1% dexamethasone eye ointment (S.A. Alcon Couvreur Pharmaceutical Co., Ltd.) was applied one time at night for one month (20). After 6 months, the frequency of prednisolone eye drops was reduced to twice/day and after one year, to once/day. Prednisolone was stopped after a period of two years.

Eyes were examined daily using a slit lamp (YZ5S digital slit lamp microscope, Shanghai Hanfei Medical Equipment Co., Ltd.) for the first 3 post-operative days, then re-examined once a week and at weekly intervals thereafter for the first month. Routine examination included uncorrected visual acuity (Snellen visual acuity chart), intraocular pressure and slit biomicroscopy examinations. Icare rebound tonometer (TAO1i; Icare Finland Oy) was selected for intraocular pressure measurement. Digital tonometry was selected if Icare couldn't get results due to cornea condition. Slit biomicroscopy was operated by the same skilled ophthalmologist.

The absorption of intraocular inflammation in the early post-operative period (one month) was assessed under a slit lamp biomicroscopy. All examinations were operated by the same ophthalmologist who performed the surgeries. The following parameters were assessed. The scoring of aqueous flare, aqueous cells, cellulose exudation and hyphema presence were graded according to the uveitis inflammation scoring criteria (21). The graft transparency, degree of graft edema and corneal neovascularization were graded according to the cornea rejection scoring criteria (22). These items included graft clarity, graft edema, neovascularization, anterior chamber flare, anterior chamber cells and exudation received between 0-4 points. Hyphema received a score of between 0-3 points. When the signs were more severe, the scores were higher. The grading systems are summarized in Table II.

S was defined as the post-operative early intraocular inflammation absorption fraction and was calculated as the difference between the post-exudation scores on the first day (T1d) and after one month (T1m; S=T1d-T1m). Additionally, ΔS was defined as the graft survival state and is the difference between the graft state on the first post-operative day (SS1d) and one month post-surgery (SS1m).

All 20 patients were followed up for 12-18 months. Visual acuity, intraocular pressure and thorough slit lamp eye exams were performed at every visit and symptoms and complications were documented.

SPSS software (version 23.0; IBM Corp) was used for statistical analysis. Data were tested for normality using the Kolmogorov-Smirnov test and provided as mean and standard deviation (SD). A paired sample t-test was used to compare the difference between T1d and T1m and the difference between SS1d and SS1m. P<0.05 was considered to indicate a statistically significant difference.

Post-operative inflammatory responses (T1d, T1m and S) and graft survival state scores (SS1d, SS1m and ΔS) are summarized in Table III. The average score of intraocular inflammation was 7.4±2.06 on the first day and 2.0±2.47 by the first post-operative month. The mean of S was 5.4±2.13. This difference was statistically significant (P<0.01). The average of SS1d was 5.3±1.56 and 3.8±1.24 for SS1m. ΔS was also statistically significant (P<0.01) Early postoperative anterior images of 2 patients are depicted in Figs. 1 and 2.

Of the 20 patients, 6 cases resulted in transplant failure. In these cases, 4 were due to rejection, a single case was due to corneal epithelial defects and a further case was caused by uncontrolled infection recurrence. Patient follow-up times and final outcomes are presented in Table IV. The survival rates of the grafts are presented in Fig. 3. At 4 months post-surgery, one patient exhibited transplant rejection leading to graft failure. Two cases exhibited failure at 8 months post-surgery, one of which was due to persistent epithelial defects. There were three cases of graft failure at 11 months post-surgery, two of which were due to rejection and one failed due to the recurrence of fungal infection.

The pre-operative visual acuity examination for all 20 patients was light perception. Post-operative visual acuity increased in 16 patients, decreased in 2 patients and 2 patients maintained pre-operative visual acuity. The specific vision changes prior to surgery and the best measured uncorrected visual acuity during the follow-up period are presented in Fig. 4. The pre-operative visual acuity of all patients was <0.05 and the optimal visual acuity after surgery was improved compared with that before surgery.

An association between early post-operative inflammatory absorption and long-term rejection based on a 12-month follow-up period of 20 patients was observed and is demonstrated in Fig. 5. In the 4 patients who exhibited transplant rejection, the early inflammatory absorption scores were relatively low, suggesting that long-term graft rejection may have a relationship with poor early inflammation control.

Post-operative complications included a persistent corneal epithelial defect that ultimately led to failure. A total of 3 cases were diagnosed with secondary fungal infections (at 6, 8 and 11 months after surgery), of which 2 cases were medically controlled and a single case resulted in graft failure. Additionally, one patient developed secondary glaucoma and lost light perception in that eye. One patient with a 9.5 mm graft had persistent epithelial defects which ultimately healed by wearing a soft bandage contact lens.