Incidence and risk factors for glaucoma development and progression after corneal transplantation

Abstract

Objective

To assess the cumulative incidence and risk factors for glaucoma development and progression within 1-2 years following corneal transplant surgery.

Design

Retrospective cohort study.

Methods

Patients undergoing penetrating keratoplasty (PK), deep anterior lamellar keratoplasty (DALK), Descemet stripping endothelial keratoplasty (DSEK), Descemet membrane endothelial keratoplasty (DMEK), Boston keratoprosthesis type I (KPro) implantation, or endothelial keratoplasty (DSEK or DMEK) under previous PK (EK under previous PK) at one academic institution with at least 1 year of follow-up were included. Primary outcome measures were cumulative incidence of glaucoma development and progression after corneal transplant, in patients without and with preoperative glaucoma, respectively. Risk factors for glaucoma development and progression were also assessed.

Results

Four hundred and thirty-one eyes of 431 patients undergoing PK (113), DALK (17), DSEK (71), DMEK (168), KPro (35) and EK under previous PK (27) with a mean follow-up of 22.9 months were analyzed. The 1-year cumulative incidence for glaucoma development and progression was 28.0% and 17.8% in patients without and with preoperative glaucoma, respectively. In a Cox proportional hazards analysis, DSEK surgery, KPro implantation, average intraocular pressure (IOP) through follow-up and postoperative IOP spikes of ≥30 mmHg were each independently associated with glaucoma development or progression (p < 0.04 for all).

Conclusions

A significant proportion of patients developed glaucoma or exhibited glaucoma progression within 1 year after corneal transplantation. Patient selection for DSEK may partly explain the higher risk for glaucoma in these patients. Postoperative IOP spikes should be minimized and may indicate the need for co-management with a glaucoma specialist.

Introduction

Glaucoma is an irreversible cause of vision loss among corneal transplant patients [1, 2]. These patients are at high risk because of the multifactorial pathophysiology, and the challenges in diagnosing, managing, and monitoring glaucoma following corneal transplantion [3–5]. Assessing the incidence and risk factors for glaucoma in corneal transplant patients is particularly important at a time of rapidly evolving corneal surgical techniques. Endothelial keratoplasty (EK), including Descemet stripping endothelial keratoplasty (DSEK) and Descemet membrane endothelial keratoplasty (DMEK), has become the most common keratoplasty procedure performed in the United States since 2012, with around 30,650 EK procedures performed in 2019. Full thickness penetrating keratoplasty (PK), by contrast, has steadily declined [6].

With increasing corneal transplantations worldwide, glaucoma has emerged as a significant post-surgical complication [5, 7]. The reported incidence and risk factors for glaucoma after corneal transplantation are highly variable depending on the type and indication for surgery [2, 8–16], but also due to inconsistent definitions of glaucoma and varying lengths of follow-up [12, 13, 17–20]. Furthermore, limited information is available on glaucoma progression after EK corneal transplantation.

To evaluate the incidence of glaucoma development and progression following either full-thickness or partial-thickness corneal transplantation, we conducted a large retrospective study in patients with at least one year of follow-up after corneal surgery. Furthermore, we assessed risk factors for glaucoma to guide the evidence-based care of these patients, including preoperative evaluation, counselling, and postoperative glaucoma monitoring.

Methods

Study design and study population

This was a retrospective cohort study comprising of all consecutive corneal transplant surgeries from April 2016 to December 2019 at Massachusetts Eye and Ear (MEE), a tertiary referral centre. The study was approved by the MEE institutional review board (IRB) in accordance with Health Insurance Portability and Accountability Act (HIPAA) regulations and adhered to the tenets set by the Declaration of Helsinki. A waiver of informed consent was granted by the IRB as the risk to the study subjects was minimal.

All subjects aged 18 years and older undergoing PK, DALK, DSEK, DMEK, KPro implantation, or EK under previous PK surgeries with at least 1 year of follow-up were included (Supplementary Fig. 1). One eye per patient was included. If both eyes of a patient were eligible, the eye with longer follow-up was assessed. If the eye underwent multiple corneal transplant surgeries, the surgery with the longest follow-up was considered, with the date of the subsequent corneal surgery as the endpoint of the follow-up period. Eyes with retinal and neuro-ophthalmic pathologies affecting visual acuity were excluded. Preoperative optic disc evaluation was not possible for some patients given corneal pathology, hence cup-to-disc ratio (CDR) measurements from slit lamp examination up to 3 months prior to and up to 1 month after the corneal transplant date were used as preoperative CDR. Data from 8 postoperative visits were collected: postoperative day 1, week 1, month 1, month 3, month 6, year 1, and from year 1.5 and year 2 when available.

Definition of glaucoma and glaucoma progression

Glaucoma was diagnosed in the presence of any of the following: an IOP ≥ 22 mm Hg at 2 consecutive visits at least 1 month apart, use of IOP-lowering medications, CDR ≥ 0.6, CDR asymmetry of >0.2, history of glaucoma surgery (including tube shunt, trabeculectomy and cyclodestructive procedures) prior to or concurrent with the corneal transplant surgery, and/or a documented diagnosis of glaucoma [8, 12–15, 21, 22]. This definition was used to classify the study population as patients with and without preoperative glaucoma. Development of glaucoma after corneal transplant was defined similarly for patients without preoperative glaucoma.

Glaucoma progression for patients with preoperative glaucoma was defined as CDR progression by ≥0.2 and/or a need for glaucoma surgery following corneal transplantation, and was not based on visual fields due to lack of reliable visual fields for a substantial portion of the study population [18, 23, 24]. Both CDR and a change in CDR were confirmed on at least two clinical exams. The need for escalation of IOP lowering therapy in patients with preoperative glaucoma was reported separately.

To identify risk factors for glaucoma development or glaucoma progression after corneal transplantation, we used a change in glaucoma status as a composite outcome consisting of glaucoma development in patients without preoperative glaucoma and glaucoma progression in patients with preoperative glaucoma.

Data collection

Data were collected and managed using the IRB-approved, HIPAA-compliant Research Electronic Data Capture (REDCap) software hosted by Harvard Medical School [25]. Demographic information and clinical data from preoperative, intraoperative, and specific postoperative time points were collected, along with any postoperative IOP spike ≥30 mm Hg [12, 26]. IOP measurements by Goldmann applanation tonometry were recorded and preferred over Tono-Pen tonometer, pneumotonometry, and digital palpation in that order.

Corticosteroid burden calculation

All corneal transplant surgeries were performed by 8 corneal surgeons at MEE, and postoperative corticosteroid management varied. To uniformly assess for postoperative topical corticosteroid burden, we assigned comparative anti-inflammatory efficacy for one drop of each steroid type to one drop of prednisolone acetate 1%. One drop of prednisolone acetate 1% was assigned an arbitrary value of 1 corticosteroid unit, and one drop of loteprednol 0.5%, fluorometholone 0.1%, prednisolone acetate 0.12% and difluprednate 0.05% were attributed values of 0.5, 0.61, 0.65 and 2 corticosteroid units, respectively [27–30]. The total corticosteroid units were calculated for each patient based on the frequency and type of corticosteroid prescribed.

Primary and secondary outcome measures

Primary outcome measure was the cumulative incidence of glaucoma development and progression after corneal transplant, in patients without and with preoperative glaucoma, respectively. Specifically, the one-year cumulative incidence was calculated per formula below.

$$\mathrm{One}\mathrm{year}\mathrm{cumulative}\mathrm{incidence}=\frac{\mathrm{Number}\mathrm{of}\mathrm{incident}\mathrm{cases}\mathrm{within}\mathrm{one}\mathrm{year}\mathrm{of}\mathrm{surgery}}{\mathrm{Total}\mathrm{number}\mathrm{of}\mathrm{cases}\mathrm{at}\mathrm{risk}\mathrm{at}\mathrm{the}\mathrm{start}\mathrm{of}\mathrm{follow\; -\; up}}$$

Secondary outcome measures included preoperative, intraoperative, and postoperative risk factors for a change in glaucoma status or the need for glaucoma surgery following corneal surgery.

Statistical analysis

Data analysis was performed using the statistical software STATA 16.1 (StataCorp LLC, College Station, Texas, USA). Categorical variables were reported as percentages and continuous variables were reported as means with standard deviations. The normality of continuous variables was assessed using the Kolmogorov–Smirnov test. Continuous variables were compared with independent samples t-test or Mann-Whitney test, contingent on the normality of the data. Tests of proportion, including the chi-square test or Fisher exact test, were used for categorical variables. Cumulative incidence of primary outcome over the entire follow-up was represented using Kaplan–Meier curve. The relation of covariates to time-to-change in glaucoma status were modelled with Cox proportional hazards survival analysis. The resulting values were expressed as hazard ratios and 95% confidence intervals (CI). All tests were two-tailed, and statistical significance was determined at p < 0.05. Where multiple comparisons were made, the family-wise error rate was controlled by applying the conservative Bonferroni correction. That is, the pre-specified alpha of 0.05 was divided by the number of pairwise tests conducted to establish revised significance thresholds (and thus reduce false positive findings). Our largest subgroup, patients who underwent DMEK, were considered our reference group unless otherwise specified.

Results

Four hundred and thirty-one eyes of 431 patients were included from a total of 952 surgeries assessed for eligibility (Supplementary Fig. 1). Of those patients, 113 underwent PK, 17 DALK, 71 DSEK, 168 DMEK, 35 KPro and 27 EK under previous PK. The most common indications were stromal or endothelial disorders (38.9%) for PK; stromal disorders (76.5%) for DALK; stromal or endothelial disorders (74.6%) for DSEK; endothelial disorders (92.9%) for DMEK; infectious pathology (31.4%) for KPro; and stromal or endothelial disorders (51.8%) for EK under previous PK (Supplementary Table 1). Overall, 35.0% patients had experienced previous graft failure. The age at surgery was 65.9 ± 14.7 years (mean ± standard deviation), with a follow-up of 22.9 ± 5.2 months (range: 12.0–35.3 months). Two hundred and forty-six patients had no preoperative glaucoma (57.1%), while 185 (42.9%) did. Reliable preoperative Humphrey visual field testing (Humphrey Visual Field [HVF] perimeter; Carl Zeiss Meditec, Dublin, CA, USA) tests, defined as fixation losses ≤33%, false-positive and false-negative rates ≤20% [31], were available for only 0.4% of patients without preoperative glaucoma and for 14.6% of patients with preoperative glaucoma.

Glaucoma development in patients without preoperative glaucoma

The 1-year cumulative incidence of glaucoma among 246 patients with no preoperative glaucoma was 28.0%, with 2 patients (0.8%) undergoing glaucoma surgery within one year. By procedure, the cumulative incidence of glaucoma within the first postoperative year was 46.0% for PK, 20.0% for DALK, 35.3% for DSEK, 18.8% for DMEK, 50.0% for KPro and 12.5% for EK under previous PK (Fig. 1A). Comparing all types of corneal transplant surgeries, a higher proportion of PK patients developed glaucoma than DMEK patients during the first year (p = 0.002). The two patients who underwent glaucoma surgery within one year postoperatively had undergone PK surgery.

Fig. 1. Incidence of glaucoma development and glaucoma progression in corneal transplant patients without and with preoperative glaucoma, respectively.

A Cumulative incidence of glaucoma development in patients without preoperative glaucoma within one year after corneal transplant surgery is depicted as a bar graph. The number of patients in each group is indicated by n. B Kaplan–Meier curves depict cumulative incidence of glaucoma development and time to development of glaucoma in patients without preoperative glaucoma after corneal transplant surgeries through follow-up. The number of patients are: PK (63), DALK (15), DSEK (17), DMEK (133), KPro (10) and EK under previous PK (8). C Cumulative incidence of glaucoma progression in patients with preoperative glaucoma within one year after corneal transplant surgery is depicted as a bar graph. The number of patients in each group is indicated by n. D Kaplan–Meier curves depict cumulative incidence of glaucoma progression and time to glaucoma progression in patients with preoperative glaucoma after corneal transplant surgeries through follow-up. The number of patients are PK (50), DALK (2), DSEK (54), DMEK (35), KPro (25) and EK under previous PK (19). Patients lost to follow-up are censored and are shown using the symbol “❙” in (B) and (D). Abbreviations: PK penetrating keratoplasty, DALK deep anterior lamellar keratoplasty, DSEK Descemet stripping endothelial keratoplasty, DMEK Descemet membrane endothelial keratoplasty, KPro Boston keratoprosthesis type I, EK under previous PK endothelial keratoplasty under previous penetrating keratoplasty.

At the final follow-up visit with an average of 23.4 ± 5.3 months after corneal surgery, 74 patients (30.1%) developed glaucoma, with 5 patients (2.0%) needing glaucoma surgery after corneal transplantation. By procedure, 46.0% PK, 26.7% DALK, 41.2% DSEK, 20.3% DMEK, 60.0% KPro and 12.5% EK under previous PK patients developed glaucoma at the final follow-up visit (p < 0.001, Table 1 and Fig. 1B). All corneal transplant surgery types had a similar follow-up duration (p = 0.46). Using the largest subgroup (DMEK patients) as a reference and comparing the types of corneal transplant surgeries, a higher proportion of PK patients developed glaucoma than DMEK patients (p = 0.002). At their final follow-up visit, 6.3% PK and 10.0% KPro patients underwent glaucoma surgery.

Table 1.

Incidence of glaucoma and glaucoma surgery at final follow-up in patients with no glaucoma before corneal transplant surgery.

Corneal surgery (n)	Length of follow-up (months)	P-value for length of follow-upa	Glaucoma after corneal transplant (% of n)	P-value for glaucoma incidencea	Pairwise comparison for glaucoma incidenceb	Glaucoma surgery after corneal transplant (% of n)
PK (63)	23.4 ± 4.6	0.46	46.0	0.001	0.002	6.3
DALK (15)	24.1 ± 6.3		26.7		>0.99	0.0
DSEK (17)	23.1 ± 5.1		41.2		0.26	0.0
DMEK (133)	23.7 ± 5.5		20.3		–	0.0
KPro (10)	21.8 ± 4.9		60.0		0.05	10.0
EK under previous PK (8)	20.1 ± 6.6		12.5		>0.99	0.0

Data are presented as mean ± standard deviation unless otherwise specified.

Significant p-values are shown in bold.

n number of patients, PK penetrating keratoplasty, DALK deep anterior lamellar keratoplasty, DSEK Descemet stripping endothelial keratoplasty, DMEK Descemet membrane endothelial keratoplasty, KPro Boston keratoprosthesis type I, EK under previous PK endothelial keratoplasty under previous penetrating keratoplasty.

^aFor comparison among all groups.

^bBonferroni corrected p-value for all corneal transplant groups compared to DMEK.

Glaucoma progression in patients with preoperative glaucoma

The 1-year cumulative incidence of glaucoma progression among the 185 patients with preoperative glaucoma was 17.8%, with 22 patients (11.9%) undergoing glaucoma surgery within one year. By procedure, the cumulative incidence of glaucoma progression within the first postoperative year was 14.0% for PK, 0% for DALK, 24.1% for DSEK, 11.4% for DMEK, 28.0% for KPro and 10.5% for EK under previous PK (Fig. 1C, p = 0.37). One year postoperatively, 10.0% PK, 14.8% DSEK, 5.7% DMEK, 24.0% KPro and 5.3% EK under previous PK patients underwent glaucoma surgery.

At the final follow-up visit with an average of 22.2 ± 4.9 months after corneal surgery, 46 patients (24.9%) exhibited glaucoma progression, and 31 patients (16.8%) underwent glaucoma surgery. By procedure, 14.0% PK, 0% DALK, 31.5% DSEK, 14.3% DMEK, 52.0% KPro and 21.0% EK under previous PK patients had glaucoma progression by final follow-up (p = 0.004, Table 2 and Fig. 1D). Similar follow-up length was available for all corneal transplant surgery types (p = 0.12). Using the largest subgroup (DMEK patients) as a reference and comparing all types of corneal transplant surgeries, a higher proportion of KPro patients showed glaucoma progression than DMEK patients (p = 0.02). At final follow-up, 10.0% PK, 20.4% DSEK, 5.7% DMEK, 40.0% KPro and 15.8% EK under previous PK patients underwent glaucoma surgery. Although not included in the definition of glaucoma progression, the proportions of patients that required escalation of glaucoma medication through follow-up were 68.0% for PK, 100.0% for DALK, 75.9% for DSEK, 54.3% for DMEK, 76.0% for KPro and 47.4% for EK under previous PK.

Table 2.

Incidence of glaucoma progression and glaucoma surgery at final follow-up in patients with glaucoma before corneal transplant surgery.

Corneal surgery (n)	Length of follow-up (months)	P-value for length of follow-upa	Glaucoma progression after corneal transplant (% of n)	P-value for glaucoma progression incidencea	Pairwise comparison for glaucoma progression incidenceb	Glaucoma surgery after corneal transplant (% of n)
PK (50)	20.8 ± 4.9	0.12	14.0	0.004	>0.99	10.0
DALK (2)	19.0 ± 8.8		0.0		>0.99	0.0
DSEK (54)	22.9 ± 5.1		31.5		>0.99	20.4
DMEK (35)	22.7 ± 4.6		14.3		–	5.7
KPro (25)	21.6 ± 5.5		52.0		0.02	40.0
EK under previous PK (19)	23.7 ± 3.0		21.0		>0.99	15.8

Data are presented as mean ± standard deviation unless otherwise specified.

Significant p-values are shown in bold.

n number of patients, PK penetrating keratoplasty, DALK deep anterior lamellar keratoplasty, DSEK Descemet stripping endothelial keratoplasty, DMEK Descemet membrane endothelial keratoplasty, KPro Boston keratoprosthesis type I, EK under previous PK endothelial keratoplasty under previous penetrating keratoplasty.

^aFor comparison among all groups.

^bBonferroni corrected p-value for all corneal transplant groups compared to DMEK.

Risk factors for change in glaucoma status

A change in glaucoma status was a composite outcome consisting of glaucoma development in patients without preoperative glaucoma and glaucoma progression in patients with preoperative glaucoma. At final follow-up, 31.9% PK, 23.5% DALK, 33.8% DSEK, 19.0% DMEK, 54.3% KPro and 18.5% EK under previous PK patients showed a change in glaucoma status.

Demographics, preoperative, intraoperative, and postoperative characteristics of patients who experienced a change in glaucoma status were compared to patients who did not (Table 3). The two groups were similar in age (overall average 65.9 ± 14.7 years), gender (overall 47.3% male) and race (overall 66.6% White, p > 0.05 for all 3 comparisons). The primary indications for corneal transplantation differed between patients who experienced a change in glaucoma status compared to those who did not. Patients with a change in glaucoma status, were more likely to have had an infectious pathology (21.7% vs 12.5%, p = 0.02) or ocular surface disease (8.3% vs. 2.6%, p = 0.007) before corneal transplantation, and less likely to have had a stromal or endothelial disorder (55.0% vs. 69.4%, p = 0.005).

Table 3.

Comparison of patients with and without a change in glaucoma status after corneal transplant surgery.

Characteristics	No change in glaucoma status after corneal transplant (n = 311)	Change in glaucoma status after corneal transplant (n = 120)	P-value
Age (years)	65.6 ± 15.1	66.8 ± 13.8	0.43
Gender (male %)	45.3	52.5	0.18
Race (White %)	64.6	71.7	0.16
Follow-up length (months)	22.9 ± 5.3	22.8 ± 4.9	0.92
Previous failed graft (%)	35.1	35.0	>0.99
Indications (%)a			0.001
· Stromal and endothelial disorders	69.4	55.0	0.005
· Infectious pathology	12.5	21.7	0.02
· Trauma	3.5	7.5	0.08
· OSD	2.6	8.3	0.007
Type of corneal transplant (%)			<0.001
· PK	24.8	30.0	0.27
· DALK	4.2	3.3	0.79
· DSEK	15.1	20.0	0.22
· DMEK	43.7	26.7	0.001
· KPro	5.1	15.8	<0.001
· EK under previous PK	7.1	4.2	0.37
Preoperative characteristics
Preoperative lens status (%)
· Phakic	37.6	37.5	0.98
· Aphakic	5.5	10.8	0.05
· PCIOL	50.8	44.2	0.22
· Others- ACIOL, scleral fixated and sulcus lens	6.1	7.5	0.60
Preoperative PAS (%)b	13.4	17.4	0.30
Preoperative uveitis (%)	3.9	5.8	0.37
Preoperative corticosteroid use (%)	43.1	42.9	0.97
Preoperative IOP (mm Hg)	14.6 ± 4.8	15.9 ± 4.6	0.01
Preoperative CCT (μm)c	697 ± 235	651 ± 155	0.15
Preoperative CDRd	0.5 ± 0.2	0.5 ± 0.2	0.74
Patients on preoperative glaucoma medications (%)	29.9	32.5	0.60
Average number of preoperative glaucoma medication	0.6 ± 1.2	0.9 ± 1.3	0.10
Intraoperative characteristics
Concurrent intraoperative surgery (%)
· CEIOL	20.6	17.5	0.47
· Anterior or pars plana vitrectomy	8.0	17.5	0.004
· Pupilloplasty or synechiolysis	11.9	15.0	0.39
Donor size (mm), mediane	7.5	7.75	0.002
· PK	8.25	8.5	0.01
· DSEK	7.75	7.75	0.86
· DMEK	7.5	7.5	0.11
Postoperative characteristics
Postoperative lens status (%)
· Phakic	18.0	15.0	0.46
· Aphakic	6.1	17.5	<0.001
· PCIOL	67.8	55.0	0.01
· Others- ACIOL, scleral fixated and sulcus lens	8.0	12.5	0.15
Average postoperative IOP (mm Hg)	15.0 ± 3.1	17.1 ± 3.1	<0.001
IOP spikes ≥30 mm Hg through follow-up (%)	15.4	45.8	<0.001
Average number of glaucoma medications through follow-up	0.6 ± 0.9	1.3 ± 1.2	<0.001
Total corticosteroid units through follow-up	201.7 ± 92.6	204.5 ± 87.0	0.77
Average CCT through follow-up (μm)	590 ± 163	623 ± 204	0.11

All data are presented as mean ± standard deviation unless otherwise specified.

Significant p-values are shown in bold.

A change of glaucoma status is defined as a composite outcome consisting of glaucoma development in patients without preoperative glaucoma and glaucoma progression in patients with preoperative glaucoma.

n number of patients, OSD ocular surface disease, PK penetrating keratoplasty, DALK deep anterior lamellar keratoplasty, DSEK Descemet stripping endothelial keratoplasty, DMEK Descemet membrane endothelial keratoplasty, KPro Boston keratoprosthesis type I, EK under previous PK endothelial keratoplasty under previous penetrating keratoplasty, PCIOL posterior chamber intraocular lens, ACIOL anterior chamber intraocular lens, PAS peripheral anterior synechiae, IOP intraocular pressure, CCT central corneal thickness, CDR cup to disc ratio, CEIOL cataract extraction and intraocular lens placement.

^aMost common indications for patients with change in glaucoma status are depicted; p-value represents comparison for all indication subcategories.

^bAvailable for 96.6% of all patients.

^cAvailable for 63.3% of all patients.

^dAvailable for 78.4% of all patients.

^eAvailable for 98.7% of PK, DALK, DSEK, DMEK and EK under previous PK patients.

Preoperatively, patients with a change in glaucoma status had a higher IOP at the preoperative visit compared to those without (15.9 ± 4.6 mm Hg vs. 14.6 ± 4.8 mm Hg, p = 0.01, Table 3). Preoperative CDR was available for 78.4% of all patients and was similar for both groups (0.5 ± 0.2 vs. 0.5 ± 0.2, p = 0.74).

Intraoperatively, a higher proportion of patients who developed a change in glaucoma status underwent concurrent anterior or pars plana vitrectomy (17.5%) compared to those without a change (8.0%, p = 0.004). Donor graft size was available for 98.7% of PK, DALK, DSEK, DMEK and EK under previous PK patients, and was significantly larger for those with a change in glaucoma status than those without (p = 0.002); specifically, PK patients with a change in glaucoma status had a larger donor graft size (median 8.5 mm) compared to those without a change in glaucoma status (median 8.25 mm, p = 0.01).

Postoperatively, more patients with a change in glaucoma status were aphakic (17.5%) and fewer were pseudophakic (55.0%) compared to those with no change in glaucoma status (6.1%, p < 0.001; 67.8%, p = 0.01; respectively). Through follow-up, average IOP (calculated as average of IOP measurements at up to 8 postoperative visits) was higher for patients who developed a change in glaucoma status than patients who did not (17.1 ± 3.1 mm Hg vs. 15.0 ± 3.1 mm Hg, p < 0.001, Table 3). A higher proportion of patients who developed a change in glaucoma status experienced an IOP spike of ≥30 mm Hg through follow-up (45.8%) compared to the patients without a change (15.4%, p < 0.001). Patients with a change in glaucoma status were on more glaucoma medications (1.3 ± 1.2 vs. 0.6 ± 0.9, p < 0.001) but did not differ in corticosteroid burden through follow-up (204.5 ± 87.0 corticosteroid units vs. 201.7 ± 92.6 corticosteroid units, p = 0.77).

Separate univariate analyses of patients without and with preoperative glaucoma displayed similar trends as the composite outcome, when evaluated for risk factors for glaucoma development (Supplementary Table 2) and glaucoma progression (Supplementary Table 3), respectively. Additionally, a univariate analysis of DALK patients was included (Supplementary Table 4) to show the clinical characteristics of this small group of patients, given the higher incidence of glaucoma (26.7% de novo) than previously reported [16].

In a Cox proportional hazards model, average IOP through follow-up (HR 1.2, CI 1.1–1.2, p < 0.001) and postoperative IOP spike of ≥30 mm Hg (HR 2.3, CI 1.4–3.7, p = 0.001) were associated with a higher relative hazard for a change in glaucoma status after corneal transplantation, while age, gender, race, indication, preoperative IOP and average postoperative corticosteroid units were not (p > 0.08 for all, Table 4). The HR for the association between surgery type and a change in glaucoma status were 1.8 for DSEK (CI 1.04–3.2, p = 0.04), and 3.5 for KPro (CI 1.6–7.6, p = 0.001), with DMEK used as the referent group for both comparisons. We considered DMEK surgery as our reference group as this was the largest group in our study population. In Cox regression analyses, segregated to evaluate risk factors for glaucoma development and glaucoma progression, PK surgery, KPro implantation and average IOP through follow-up were associated with glaucoma development in patients without preoperative glaucoma (Supplementary Table 2), while KPro implantation and postoperative IOP spike ≥30 mm Hg were associated with glaucoma progression in patients with preoperative glaucoma (Supplementary Table 3). Although not significant in the multivariate analysis, fewer PK patients with preoperative glaucoma experienced progression (p = 0.04 in univariate analysis, Supplementary Table 3).

Table 4.

Cox proportional hazards analysis to evaluate risk factors for change in glaucoma status after corneal transplant surgery.

Variable	Hazards ratio	95% Lower bound	95% Upper bound	P-value
Age	1.00	0.99	1.02	0.48
Gender (ref = male)	0.8	0.5	1.2	0.28
Race (ref = White)	0.7	0.5	1.1	0.14
Indication (ref = stromal and endothelial pathologies)
· Infectious pathology	1.2	0.7	2.3	0.53
· Trauma	1.1	0.5	2.6	0.76
· OSD	1.4	0.6	3.4	0.40
· Others	0.5	0.2	1.1	0.08
Surgery performed (ref=DMEK)
· PK	1.4	0.7	2.7	0.33
· DALK	0.9	0.3	3.0	0.91
· DSEK	1.8	1.04	3.2	0.04
· KPro	3.5	1.6	7.6	0.001
· EK under previous PK	0.8	0.3	2.3	0.74
Preoperative IOP	1.02	0.98	1.1	0.36
Average postoperative IOP	1.2	1.1	1.2	<0.001
Postoperative IOP spike ≥30 mm Hg (ref=no)	2.3	1.4	3.7	0.001
Average postoperative corticosteroid units	1.00	0.99	1.00	>0.99

Significant p-values are shown in bold.

OSD ocular surface disease, PK penetrating keratoplasty, DALK deep anterior lamellar keratoplasty, DSEK Descemet stripping endothelial keratoplasty, DMEK Descemet membrane endothelial keratoplasty, KPro Boston keratoprosthesis type I, EK under previous PK endothelial keratoplasty under previous penetrating keratoplasty, IOP intraocular pressure.

Overall, 36 patients (8.3%) needed glaucoma surgery after corneal transplantation by final follow up, specifically by surgery: 8.0% PK, 0% DALK, 15.5% DSEK, 1.2% DMEK, 31.4% KPro and 11.1% EK under previous PK. Univariate analysis (Supplementary Table 5) and Cox proportional hazards model (Supplementary Table 5) yielded the following risk factors for post-transplant glaucoma surgery: male gender, non-White race, DSEK surgery, KPro implantation and postoperative IOP spike ≥30 mm Hg.

A subgroup analysis was performed comparing DSEK patients to those who underwent DMEK surgery, as DSEK surgery was identified as a risk factor for glaucoma status change and need for glaucoma surgery. DSEK patients were older than DMEK patients and a higher proportion of them were non-White (Supplementary Table 6). Univariate analysis showed differences in preoperative characteristics, such as more patients with previous failed grafts, uveitis, peripheral anterior synechiae (PAS), aphakia, and a higher proportion of patients who had undergone preoperative glaucoma surgeries in the DSEK group. During the postoperative period, a higher proportion of DSEK patients experienced IOP spikes of ≥30 mm Hg, had thicker corneas (667 ± 132 µm vs. 533 ± 47 µm, p < 0.001) and experienced a higher corticosteroid burden compared to DMEK patients.

Discussion

In this large retrospective study, we showed that the 1-year cumulative incidence of glaucoma development and glaucoma progression were 28.0% and 17.8% for corneal transplant patients without and with preoperative glaucoma, respectively. At a mean follow-up of 22.9 months after corneal transplantation, 30.1% of patients without preoperative glaucoma developed glaucoma, while 24.9% of patients with preoperative glaucoma exhibited glaucoma progression. For newly diagnosed glaucoma, other recent studies with comparable follow-up times have reported incidence of glaucoma as 8.7-34% after PK, 0–4.5% after DALK, 11.9–36% after DSEK, 2.7–6.6% after DMEK and 21.4–54.2% after KPro surgery [8–10, 12–16, 24, 32]. In our study, 46.0% PK, 26.7% DALK, 41.2% DSEK, 20.3% DMEK, 60.0% KPro and 12.5% EK under previous PK patients without preoperative glaucoma developed glaucoma. While our results are mostly consistent with the literature, we reported a higher incidence of glaucoma development in DALK patients, probably due to our small subgroup size of 15 patients and our definition of glaucoma, which included an increase in optic nerve CDR [8, 15, 21, 22], in addition to criteria previously used by others, namely, IOP elevation requiring glaucoma medication or surgical intervention [12–14]. Given glaucoma is an optic neuropathy, we felt that its definition in our study should include the optic nerve status [33]. Of note, only 1 DALK patient developed increased CDR after surgery (6.7%), which is comparable to a previous report with this definition of glaucoma [16]. For glaucoma progression in patients with preoperative glaucoma, previous studies reported the incidence as 5.7–68% for PK, 33.3% for DALK, 4–63% for DSEK, 2.5% for DMEK and 21-45% in KPro patients [9, 11, 16, 34–37]. We reported 14.0% PK, 0% DALK, 31.5% DSEK, 14.3% DMEK, 52.0% KPro and 21.0% EK under previous PK patients with preoperative glaucoma exhibited glaucoma progression after the corneal transplant surgery. While most of our results were similar to the prior literature, we reported a higher incidence of glaucoma progression in DMEK patients. This may be due to the difference in our definition of glaucoma progression from previous studies: for reason listed above, we included optic nerve changes, but did not include escalation in glaucoma medication as glaucoma progression, although we reported these results separately. Interestingly, few of our PK patients with prior glaucoma experienced glaucoma progression (14.0%), while more of them without preoperative glaucoma developed glaucoma (46.0%) after surgery. This may be explained by patient selection bias, as more DSEK patients had preoperative glaucoma (76.1%) compared to PK patients (41.6%, p < 0.001), probably because more patients with preoperative glaucoma were offered DSEK surgery due to the well-known risk of glaucoma after a PK surgery [38]. Moreover, PK patients with preoperative glaucoma were significantly younger than DSEK patients with preoperative glaucoma (64.4 ± 14.5 years vs 74.2 ± 10.9 years, p < 0.001); older age has been determined as a risk factor for glaucoma progression previously [39]. Risk factors associated with a change in glaucoma status were also assessed, defined as a composite of glaucoma development in glaucoma-naïve patients and glaucoma progression in patients with preoperative glaucoma; similar results were obtained when we analyzed the risk factors for the two groups separately (Supplementary Tables 2 and 3). Similar to previous literature, we identified risk factors, such as indication of cornea surgery and preoperative IOP, in our univariate analysis [11, 13, 40, 41]. In the multivariate analysis, we showed that postoperative IOP spikes of ≥30 mm Hg, a higher average postoperative IOP, along with DSEK and KPro surgery, were significantly associated with an increased relative hazard of glaucoma. Additionally, we found that male gender, non-White race, DSEK surgery, KPro implantation and postoperative IOP spike of ≥30 mm Hg were associated with glaucoma surgery after corneal transplant. These findings were similar to previous studies, which have reported preoperative glaucoma, male gender, and Black race to be associated with glaucoma surgery after corneal transplantation [42]. The risk factors for glaucoma and glaucoma surgery identified by us and others may guide preoperative counselling and early referral for co-management with a glaucoma specialist to prevent glaucomatous vision loss in patients undergoing corneal transplantation.

One interesting finding from our study is the high incidence of change in glaucoma status and need for glaucoma surgery after DSEK surgery when compared to DMEK. At final follow-up, 41.2% of DSEK and 20.3% of DMEK patients developed glaucoma. In addition, 31.5% of DSEK patients with preoperative glaucoma demonstrated glaucoma progression, with 20.4% undergoing glaucoma surgery and 75.9% requiring escalation of glaucoma medication at a mean follow-up of 22.9 months. Prior studies have reported that 4.5-27% of DSEK patients without preoperative glaucoma developed post-DSEK glaucoma or had a postoperative IOP elevation requiring glaucoma medication or surgical intervention with 0-0.3% needing glaucoma surgery [12, 19, 43], while 20-47.4% of DSEK patients with preoperative glaucoma required escalation of glaucoma medication and 5.8-19% needed glaucoma surgery [19, 43, 44]. The high incidence of glaucoma development, need for additional glaucoma medication and glaucoma surgery postoperatively in DSEK patients in our study may be related to patient selection rather than the surgery itself (Supplementary Table 6). MEE is a tertiary referral centre, and many patients have complex anterior segment pathology prior to corneal surgery. Nearly 80% of our DSEK patients had preoperative glaucoma, and 50.7% had a history of glaucoma surgery. Cornea surgeons at our centre have a preference for DSEK over DMEK in patients with prior glaucoma surgery [45, 46]. Compared to DMEK patients, our DSEK population was older and had a higher proportion of non-White patients, which are risk factors for glaucoma [47, 48]. Compared to DMEK patients, DSEK patients were more likely to have had previous graft failure, history of uveitis, aphakic lens status, and PAS, all of which have been associated with glaucoma [49–51]. Finally, we found that DSEK patients had thicker average CCT through follow-up suggesting that the average IOP may have been underestimated in these patients given possible corneal oedema [52].

A large proportion of KPro patients developed glaucoma (60%) or exhibited glaucoma progression (52%) after KPro implantation in our cohort. The high incidence is similar to the previously reported incidence of glaucoma (24%-54.2%) and glaucoma progression (43.7%-85.7%) in KPro patients [11, 24, 32, 53]. The mechanism of glaucoma and glaucoma progression in KPro patients is multifactorial, and may involve iridocorneal adhesions following progressive shallowing of the anterior chamber, clogging of the trabecular meshwork with inflammatory debris and inflammatory cytokine-mediated damage to the retinal ganglion cells [54, 55]. In addition, diagnosis and monitoring of increased IOP and glaucoma can be very challenging in KPro patients due to inability to accurately measure IOP, difficulty in obtaining visual fields, fundus photography and other imaging tests of the optic nerve head.

Limitations to our study include those inherent to a retrospective design and a relatively small sample size for some corneal transplant types, such as the DALK group. CDR was not reported at each visit, due to corneal pathology, postoperative oedema, and/or photophobia, limiting examination of the optic nerve. This, however, is consistent with practice guidelines for patients with corneal oedema and opacification [56]. We included CDR as part of the definition of glaucoma but could not use more objective measures, such as optical coherence tomography and HVF, which were not consistently available in this patient population. Lastly, our results should be interpreted in light of the study population sourced from a tertiary academic centre, where patients may be different from the general community.

In conclusion, corneal transplant surgeries are associated with glaucoma development and progression, with DSEK and KPro surgeries at particularly high risk. As lamellar transplants have now become the most common form of corneal transplant procedures [6], it is essential to counsel patients preoperatively regarding the risk of glaucoma. Postoperative IOP spikes should be minimized and may indicate the need for co-management with a glaucoma specialist. Our study provides important clinical parameters for managing glaucoma in high-risk corneal transplant patients, who may require multidisciplinary care to reduce irreversible vision loss.

Supplementary information is available at nature.com/eye

Summary

What was known before

• Glaucoma leads to irreversible vision loss in corneal transplant patients.

• Reported incidence and risk factors are highly variable depending on the type of corneal transplant, indication of the surgery but also due to inconsistent definitions of glaucoma and varying lengths of follow-up.

• Limited information is available on glaucoma after lamellar corneal transplantation.

What this study adds

• Based on a definition of glaucoma which includes IOP elevation, glaucoma therapy and optic nerve changes, we reported the 1-year cumulative incidence for glaucoma development and progression after various types of corneal surgery to be 28.0% and 17.8% in patients without and with preoperative glaucoma, respectively.

• The incidence of glaucoma development and progression in Descemet stripping endothelial keratoplasty (DSEK) patients after an average follow-up of 23 months was 41.2% and 31.5%, respectively. Patient selection for DSEK may partly explain the higher risk for glaucoma in these patients.

• Risk factors for glaucoma after corneal surgery include DSEK and KPro surgery, high average IOP through follow-up and postoperative IOP spike ≥30 mm Hg. Postoperative IOP elevation and IOP spikes are important clinical parameters which may indicate the need for co-management with a glaucoma specialist in order to prevent irreversible vision loss in patients with corneal transplantation.

Author contributions

CS: Conceptualization, methodology, formal analysis, investigation, data curation, writing- original draft, visualization. ECD: Conceptualization, methodology, investigation, resources, writing- review & editing. LU: Conceptualization, methodology, formal analysis, writing- review & editing. JC: Conceptualization, methodology, investigation, resources, writing- review & editing, supervision, funding acquisition. JBC: Conceptualization, methodology, investigation, resources, writing- review & editing. UVJ: Conceptualization, methodology, investigation, resources, writing- review & editing. EIP: Conceptualization, methodology, investigation, resources, writing- review & editing. RP: Conceptualization, methodology, investigation, resources, writing- review & editing. HNS: Conceptualization, methodology, investigation, resources, writing- review & editing. JY: Conceptualization, methodology, investigation, resources, writing- review & editing. LQS: Conceptualization, methodology, investigation, resources, writing- original draft, supervision, project administration, funding acquisition.

Funding

Boston Keratoprosthesis Fund, Massachusetts Eye and Ear, Boston, MA.

Data availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests

CS is supported by the Boston Keratoprosthesis Fund, Massachusetts Eye and Ear, Boston. LU is supported in part by the Dozoretz Family Private Foundation. JC is a consultant for the US Food and Drug Administration. JBC is a consultant for ORA System, TherOptix and Fontana and, has intellectual property and equity interest in TherOptix and Fontana. RP is a consultant for Sanofi Genzyme and receives royalties from Elsevier. JY is a consultant for Kera Therapeutics and has equity interest in Kera Therapeutics. LQC received research support from Topcon and is a consultant for FireCyte Therapeutics and AbbVie. ECD, EIP, and HNS declare no potential competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41433-022-02299-6.