Excessive corneal endothelial single-cell loss following endothelial injuries

Abstract

PURPOSE:

The purpose of this study was to investigate corneal endothelial changes following common clinical endothelial injury scenarios in order to uncover mechanisms underlying unexplained chronic corneal endothelial wound healing.

MATERIALS AND METHODS:

This cross-sectional study included patients with endothelial injuries from three common scenarios: postcataract surgery, corneal dystrophies, and penetrating injuries. Noncontact specular microscopy was used to capture images from five distinct corneal regions. Endothelial cell density (ECD), coefficient of variation (CV), and percentage of hexagonal cells (HEX) were assessed. All endothelial photographs were also reviewed. Statistical analysis was performed to compare injured and noninjured eyes.

RESULTS:

Seventy-seven patients were enrolled, with a mean age of 64 years (48 females, 29 males). The mean central ECD was 2138.91 ± 869.34 cells/mm² in postcataract surgery eyes, 1999.48 ± 763.91 cells/mm² in endothelial dystrophy eyes, and 1854.86 ± 551.85 cells/mm² in trauma cases. While most parameters showed no significant differences, postcataract surgery eyes exhibited a significant increase in CV value in the upper and temporal regions (P < 0.05). Unexpectedly, stochastic single-cell loss was observed in 42.86% of patients, continuing up to two years postinjury. This loss was significantly higher compared to uninjured eyes (P = 0.00005), suggesting that excessive single-cell loss occurs well beyond the expected wound healing period.

CONCLUSION:

We identified accelerated stochastic single-cell loss in the corneal endothelium following primary injuries, persisting well beyond the expected wound healing period, a phenomenon that has not been previously highlighted. This finding offers a potential explanation for the chronic endothelial cell loss following a primary injury.

Introduction

Corneal endothelial dysfunction is one of the leading causes of corneal transplantation, underscored by its prevalence in over 50% of such cases.[1] The corneal endothelium is a single-layered cell lining the inner side of the cornea, which responds to corneal nutrition and the regulation of corneal hydration for keeping the cornea in a relatively dehydrated status for maintaining the cornea in a transparent status.[2] Unlike corneal epithelium, corneal endothelium does not replicate in vivo at normal physiological status, even though it possesses the capability of proliferation.[2] Under normal physiological conditions, the adult human corneal endothelial cell density (ECD) is approximately 2500-3000 cells/mm²[3] and decreases 0.6% annually by age.[4] Any injury or disease that causes additional endothelial cell damage can lead to an irreversible decompensation of the cornea. When the ECD drops below a critical threshold of approximately 500 cells/mm²,[5] it results in the swelling of the cornea, loss of corneal transparency, and consequential vision loss. The corneal endothelium’s nonproliferative nature and limited regenerative capacity cause a significant challenge for endothelial regeneration. Although some pharmaceutical approaches have been developed recently, currently, the mainstay for treating corneal endothelial decompensation is still allogeneic transplantation.[6]

The comprehensive mechanism of corneal endothelium wound healing remains unclear.[2] Unlike corneal epithelium, the endothelial cells are mostly arrested in the G1 phase, and heal wounds by cell enlargement and migration of adjacent cells, but not by proliferation.[7] Therefore, it is believed that when the damaged area is too large to be completely covered, the residual cell density drops beyond the threshold to maintain normal watering pumping function. The cornea then decompensates, becoming swollen and not transparent, leading to the loss of vision. Pseudophakic bullous keratopathy (PBK), Fuchs’ endothelial dystrophy (FED), and trauma are the common clinically encountered endothelial diseases that cause irreversible endothelial decompensation.[1] PBK refers to irreversible corneal endothelial dysfunction following cataract extraction and intraocular lens implantation surgery.[8] The prevalence of PBK is approximately 1%.[9] In PBK, the primary cause of endothelial cell loss is surgical trauma during phacoemulsification. This trauma is attributed to multiple factors, including thermal injury from elevated local temperatures near the phacoemulsification probe, turbulent fluid flow, mechanical stress from ultrasound energy, and the generation of free radicals and oxidative stress during the procedure.[8] FED is the most common form of endothelial dystrophy which is characterized by the endothelial cell loss and development of guttae, the excrescences of Descemet’s membrane. FED progresses slowly within time and will expand to significant endothelial cell loss, subsequential loss of corneal deturgescence, and irreversible corneal edema.[10] The prevalence of FED varies greatly, with an estimated 7.33% globally.[11] In some clinical scenarios of corneal endothelial injury, we did observe that the corneal endothelial cell loss progresses even if the causative incident did not proceed. The onset of bullous keratopathy following cataract surgery and argon laser iridotomy (ALI) was reported at mean 148.4 and 107 months, respectively.[12] The exact mechanism for the progression of endothelial cell loss beyond normal aging loss following primary insult remains uncertain. Therefore, in this observational study, we revisit the endothelial cell loss in three common causes of corneal endothelial injury, including postcataract surgery, endothelial dystrophy, and penetrating corneal injury, to identify the potential insights into corneal endothelial wound healing.

Materials and Methods

Study design

This cross-sectional observational study received approval from the Institute Review Board of Chang Gung Memorial Hospital, Linkou, Taiwan (approval number: 202200118B0); approval date: May 3, 2022). Written informed consent was obtained from all recruited patients prior to data collection. We included patients with three common causes of endothelial injury, including postcataract surgery, corneal endothelial dystrophies, and penetrating corneal with record at Chang Gung Memorial Hospital between March 2018 and July 2023. Basic data on patient demographics, diagnosis, and treatment history were reviewed and collected. Subjects diagnosed with active corneal diseases other than endothelial dystrophies, as well as those who had undergone corneal transplantation, were excluded. The patients with severe corneal edema observed under slit-lamp biomicroscopy that confounded specular microscopic examinations were also excluded. The noncontact in vivo specular microscopy (CEM-530, Nidek, Gamagori, Japan) was employed for endothelial cell morphological analysis. The endothelial photographs were captured at five distinct regions of the cornea for each patient using a standard magnification of × 400 [Figure 1]. These regions included the central point – with the fixation light aimed directly at the center – and four paracentral points (upper, lower, nasal, and temporal). For the paracentral points, the fixation light targeted a paracentral zone with a diameter of 1.3 mm at specific angular positions. In the right eye, these positions were at 90°, 270°, 0°, and 180°, while in the left eye, they were at 90°, 270°, 180°, and 0°, respectively. The obtained images were processed using the manufacturer-provided software, which calculated morphometric parameters including cell density (ECD), coefficient of variation (CV), and percentage of hexagonal cells (HEX) through the inbuilt image analysis system. Additionally, stochastic single-cell loss, identified as endothelial single-cell dropout, was detected through the review of endothelial photographs.

Figure 1.

Representative corneal endothelial images captured at five distinct locations: the central point and four paracentral points at 0°, 90°, 180°, and 270° (×400)

Statistical analysis

All statistical analyses in this study were conducted using SPSS version 29.0 (SPSS Inc., Chicago, IL, USA). Descriptive statistics were applied to the basic data, including age, sex, and laterality of the eye across the three groups: postcataract surgery, endothelial dystrophy, and trauma. The endothelial morphometric parameters included ECD, CV, and HEX from 5 distinct regions for each case, which were also compared between the injured eye and the noninjured fellow eye, using the Wilcoxon signed rank test, two-sided. Fisher’s exact test was used to compare the presence of stochastic single-cell loss between healthy and injured eyes.

Results

In this cross-sectional observational study, we enrolled patients with three common causes of endothelial injury, including postcataract surgery, corneal endothelial dystrophies, and penetrating corneal injury. We included 77 patients in this observational study. The mean age was 64 years, with 48 females and 29 males. The endothelial photographs taken by specular microscopy that produced unclear images, where endothelial morphology could not be detected or endothelial parameters could not be analyzed, were excluded from the statistical analysis.

We collected effective endothelial data from 32 eyes who had undergone cataract surgery without having endothelial dystrophy, 32 eyes with endothelial dystrophy who had not undergone cataract surgery, 14 eyes with endothelial dystrophy who had received cataract surgery, and 7 eyes with penetrating corneal injuries. The mean duration from primary injury to examination, using the central region as representative, was 26.64 ± 55.69 weeks, ranging from 1 day to 249.29 weeks for cataract surgery cases, and 28.07 ± 29.90 weeks, ranging from 1 day to 77.57 weeks for trauma cases [Table 1].

Table 1.

Selected endothelial parameters from five distinct corneal regions following three common endothelial injuries

Injury type	Location	n	ECD (cells/mm2)	CV (%)	HEX (%)	Duration (weeks)
Endothelial dystrophy	Central	46	1999.48±763.91	36.74±14.45	56.28±22.36	N/A
	Superior	29	2157.31±682.3	33.10±8.32	60.62±10.08	N/A
	Inferior	30	2445.03±519.45	34.3±8.78	61.67±15.67	N/A
	Nasal	30	2366.43±534.42	32.53±7.67	62.8±12.61	N/A
	Temporal	27	2033.93±766.8	34.85±11.75	56.85±21.25	N/A
Postcataract surgery	Central	46	2138.91±869.34	34.98±14.42	58.63±18.20	26.64±55.69
	Superior	41	2192.12±866.3	33.27±8.40	58.22±15.36	26.37±59.17
	Inferior	42	2395.52±1001.5	32.02±12.42	58.67±15.19	26.58±58.32
	Nasal	43	2384.53±704.05	32.42±8.36	62.23±10.38	25.89±57.63
	Temporal	42	2205.12±798.02	35.69±12.02	62.60±14.50	26.53±58.34
Corneal penetrating injury	Central	7	1854.86±551.85	27.57±4.20	61.71±27.73	28.07±29.90
	Superior	5	2557.0±273.57	29.0±3.0	67.8±7.05	16.23±21.95
	Inferior	3	2484.33±167.31	28.67±7.09	67.33±7.23	22.76±28.35
	Nasal	4	2526.5±267.13	29.5±6.24	66.25±3.95	18.68±24.55
	Temporal	5	1801.0±707.57	36.2±11.41	46.8±27.37	13.89±23.45

Duration=Duration from primary injury to examination. ECD=Endothelial cell count, CV=Coefficient of variant, HEX=Percentage of hexagonal cells, N=Number of photographs, N/A=Not available

The mean ECD at the center was 2138.91 ± 869.34 cells/mm² in the postcataract surgery group (n = 46), 1999.48 ± 763.91 cells/mm² in the FED group (n = 46), and 1854.86 ± 551.85 cells/mm² in the trauma group (n = 7). The endothelial morphometric parameters are listed in Table 1. We compared the eyes postcataract surgery with their uninjured fellow eyes and found no significant differences in most parameters of the endothelium after injury at five distinct regions of the cornea [Table 2]. A significant increase in CV value was observed in the upper and temporal regions of postcataract surgery eyes (P < 0.05). Although still within the normal range, we believe this increase is due to endothelial cell loss and the subsequent wound-healing process, where neighboring cells migrate and enlarge to cover the injured area near the sites of the surgical incision and phacoemulsification probe. This cellular response leads to greater variation in cell size, resulting in a higher CV value. Due to the small sample size of 7 trauma eyes and the large variation within the group, we did not perform statistical comparisons.

Table 2.

Comparisons of selected endothelial parameters from five distinct corneal regions following cataract surgery and in the fellow eyes

Location	Central		P	Superior		P	Inferior
	Nonoperated eye	Operated eye		Nonoperated eye	Operated eye		Nonoperated eye	Operated eye
n	24			22			21
Duration (weeks)	7.58±11.53			5.61±5.20			5.66±5.36
ECD (cells/mm2)	2506.63±345.12	2473.5±455.82	0.152	2500.77±274.87	2514.50±368.08	1	2607.05±312.61	2603.52±561.53
CV (%)	30.0±6.50	30.46±8.53	0.286	29.55±6.01	33.36±6.86	0.019	30.05±4.34	32.10±6.27
HEX (%)	66.5±8.76	65.63±15.73	0.839	68.5±7.04	64.05±8.85	0.078	65.57±6.45	63.90±9.21
Location	P	Nasal			P	Temporal		P
		Nonoperated eye		Operated eye		Nonoperated eye	Operated eye
n		22				22
Duration (weeks)		5.57±5.23				5.50±5.27
ECD (cells/mm2)	0.664		2589.18±313.46	2593.14±371.50	0.548	2432.09±498.70	2463.09±396.34	0.57
CV (%)	0.167		29.09±5.62	31.91±6.63	0.081	28.95±4.60	33.95±8.72	<0.001
HEX (%)	1		67.32±6.90	64.86±8.65	0.413	67.36±5.66	67.59±7.02	0.702

P value by Wilcoxon signed ranks test between operated group and nonoperated fellow group. ECD=Endothelial cell density, CV=Coefficient of variation, HEX=Percentage of hexagonal cells

Unexpectedly, after reviewing all endothelial photographs, we observed a distinct pattern of increasing stochastic single-cell loss in the injured corneas (42.86% of patients; 29.51% of injured eyes) across all three endothelial injury scenarios. Specifically, this included 35.29% of patients (30.61% of eyes) who had undergone cataract surgery without having endothelial dystrophy, 50% of patients (27.78% of eyes) with endothelial dystrophy who had not undergone cataract surgery, 43.75% of patients (25% of eyes) with endothelial dystrophy who had received cataract surgery, and 44.44% of patients (44.44% of eyes) with penetrating corneal injuries. This finding was evident from day 1 up to 2 years following the primary injury [Figure 2]. A summary of the observed single-cell loss data is provided in Table 3.

Figure 2.

Representative endothelial photographs captured by specular microscopy, showing single-cell loss observed across all three injury scenarios. (a) Postcataract surgery, at the paracentral nasal point, 2 years after surgery; (b) Fuchs’ endothelial dystrophy; (c) Penetrating corneal injury, at the paracentral inferior point, 5 weeks postinjury (white arrows indicate single endothelial cell loss; yellow arrow indicates guttae)

Table 3.

Incidence of single-cell loss observed in three common types of endothelial injuries

Injury type	Patients	Percentage	P	Eyes	Percentage	P
Postcataract surgery	12/34	35.29	0.002	15/49	30.61	0.004
FED without cataract surgery	9/19	50	<0.001	10/36	27.78	0.01
FED with cataract surgery	7/16	43.75	<0.001	7/28	25	0.022
Corneal penetrating injury	4/9	44.44	0.002	4/9	44.44	0.002
Injured total	33/78	42.31	<0.001	36/122	29.51	0.003
Normal control	2/32	6.25		2/32	6.25

P value by sample t-test between each injury group and normal control group. FED=Fuchs’ endothelial dystrophy

In contrast, only two instances of single-cell loss 6.25%), each occurring in a single endothelial microscopy photograph, were observed in the uninjured fellow eyes of cataract surgery and trauma cases. Although the selected regions may not fully represent the response of the entire endothelium to injury, we nonetheless highlighted a phenomenon of excessive stochastic single-cell loss occurring after the primary injury, even well beyond the expected wound healing period.

Discussion

In our observational study, we investigated endothelial morphology following three common endothelial injuries: posterior-cataract surgery, FED, and trauma. We discovered an interesting new phenomenon, where excessive, random single-cell loss was consistently observed in injured corneas across different types of endothelial injuries (42.86% of patients; 29.51% of injured eyes), which was significantly less frequent in normal fellow eyes. Apart from corneal endothelial dystrophy, although the primary etiological factors differ, a common observation is the persisting disease progression that exceeds the normal aging changes, even after the main pathological process ceases.

The exact mechanism of corneal endothelium wound healing is still unclarified.[13] It remains unclear how a primary endothelial injury leads to a chronic progression of endothelial cell loss, eventually resulting in irreversible endothelial decompensation requiring transplantation. Chronic endothelial cell loss following routine cataract surgery is an example. Chronic progressive endothelial cell loss following cataract surgery has been widely reported.[14,15,16] The majority of studies evaluate the endothelium within 1 year postoperatively, with varying reported rates (2.3%–18.19%).[14,17,18] The preoperative patient condition and the surgical techniques all contribute to the large variation.[14,17] Despite the modern refinement of cataract surgery, a 2.06% endothelial cell loss rate per year[15] following uncomplicated cataract surgery, which significantly outraces physiological aging (0.6%),[19] has still been reported in a long-term (10 years) follow-up study. Moreover, another 7-year prospective study reported a total 23.47% endothelial cell loss, with an affirmed progression over time.[14]

The endothelial cell loss following cataract surgery has been proposed in two stages. The immediate postoperative endothelial cell loss is believed to be directly related to surgical procedures:[20] thermal and mechanical damages caused by phacoemulsification, contact with lens fragments during surgery, and inflammation are the main reasons for immediate endothelial cell loss following cataract surgery.[21,22,23,24] After the initial endothelial wounding is morphometrically stabilized 3 months following cataract surgery,[25] the continued and accelerated long-term endothelial cell loss has been hypothesized to be related to subclinical inflammation, decreased innervation, loss of vitreous, and possibly a tendency for endothelial remodeling – that has not been confirmed until recently.[15,16] In the 10-year follow-up study, nuclear firmness and early postoperative corneal edema are proposed as predictive factors for long-term endothelial cell loss.[15] However, their findings are to an extent contradicted by the late endothelial remodeling hypothesis, which states there is no difference in the morphological indices between pre- and postoperative endothelium.[15] Moreover, in the 7-year follow-up study, Lundberg confirmed the trend of chronic endothelial cell loss.[14] Interestingly, the late endothelial cell loss is less in corneas with significant immediate postoperative corneal edema, and the total cell loss is similar regardless of the severity of early corneal edema. It therefore implies a possibility that the primary endothelial injury per se initiates the long-term endothelial cell loss, regardless of the severity of early corneal edema.[14]

Another clinical example of significant chronic corneal endothelial cell loss not included in this study is the ALI, a treatment for angle-closure glaucoma that can result in bullous keratopathy, which has an incidence of corneal edema at approximately 1.8% and up to 20% in corneal transplant cases in Japan, typically manifesting 5.5–7.4 years post-ALI.[26,27,28] While several mechanisms have been proposed,[26,27,28] an unidentified chronic damaging process likely drives late-onset corneal edema.

In this observational study, we revisited the endothelial morphometric changes following three common endothelial injuries. We observed an interesting new phenomenon of excessive stochastic single-cell loss in the injured cornea (42.86% of patients; 29.51% of injured eyes), consistently across all three common clinical endothelial injury scenarios [Figure 2]. The cell loss occurred at regions distant from the main lesion [Figure 2a] as well as near the guttae [Figure 2b]. It could be observed as late as 2 years following primary injury insult [Figure 2c]. To our knowledge, it is the first report that highlighted this morphological finding of the endothelium following primary endothelial injuries. We hypothesize that the primary endothelial wound healing process might somehow induce secondary cell loss outside the primary wound during the initial healing process. Stochastic loss of healthy, possibly migrating, endothelial cells can occur during this process. Moreover, the secondary single-cell loss could potentially become another new wound, initiating further cascades of endothelial injury. This cascade of excessive cell loss might result in a chronic, progressive loss of endothelial cells, surpassing what would be expected from normal aging and extending well beyond the expected wound healing period.

When the generally accepted mechanisms of endothelial wound healing, including migration and enlargement of neighboring cells, and possibly, the induction of limited proliferative capacity of endothelium, all together fail to compensate the primary and secondary cell loss, an accelerated, chronic cell loss may occur, eventually leading to decompensation of endothelium. Further study will be conducted to clarify the mechanism for excessive stochastic cell loss following primary endothelial injury. The major limitation of our study is the lack of longitudinal follow-up data, as this is only an observational study. Additionally, although we captured images from five different regions of the endothelium, it will still be beneficial to have data covering a larger area to more comprehensively observe the wound healing response across the entire endothelium.

Conclusion

Our study highlighted a novel phenomenon of increasing stochastic single endothelial cell loss following common primary endothelial injuries, which continues well beyond the expected healing period. This finding offers insights into the unexplained chronic endothelial cell loss following an initial injury. Future research will aim to investigate the underlying mechanisms and potentially offer valuable information to improve the treatment strategies for corneal endothelial dysfunction.

Data availability statement

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

Conflicts of interest

The authors declare that there are no conflicts of interests of this paper.

Funding Statement

Nil.