Corneal Endothelial Cell Loss From Anterior Chamber Paracentesis in a Porcine Model

Abstract

This study aimed to develop a porcine eye model used to predict corneal endothelial cell loss (ECL) associated with anterior chamber paracentesis (ACP) in a healthy adult human cornea. To assess the average wound area and ECL created by needle punctures, a 27-gauge (27G) needle was inserted at the limbus of a porcine eye through the clear cornea and into the anterior chamber. Needle-punctured areas were immediately collected, stained, and photographed with a digital light microscope. The wound areas were quantified in square millimeters (mm²) and then extrapolated to predict ECL in a healthy adult human cornea. The average wound area of the puncture sites was 0.274 ± 0.122 mm². The needle punctures created a larger area of ECL than the observed cross-sectional area of the needle (0.12 mm²). Extrapolating these data to the ECL that would occur in healthy adult human corneas, each 27G corneal needle puncture would damage 685 cells, or 0.21% of the corneal endothelial cell layer. The predicted wound areas for 28G, 30G, and 33G needles were 0.234 mm², 0.163 mm², and 0.072 mm², respectively. The predicted cell loss for a 28G, 30G, and 33G needle stick corresponded to a loss of 586 cells (0.18% of the corneal endothelium), 407 cells (0.13% of the corneal endothelium), and 181 cells (0.06% of the corneal endothelium), respectively. Frequent ACPs following intravitreal injections may be associated with clinically significant ECL; thus, caution is advised, particularly in patients with compromised corneas and low endothelial cell counts.

Introduction

Anterior chamber paracentesis (ACP) allows for diagnostic sampling, immediate reduction of elevated intraocular pressure (IOP) in acute primary angle closure, and management of elevated IOP secondary to intravitreally administered medication.^1,2 ACP before or after an injection of intravitreal (IVT) medication may be required due to elevated IOP associated with transient central retinal artery occlusion with vision loss and ocular pain. Monitoring of optic nerve head perfusion is in the product labels. ³ The consequences of frequent IOP spikes, especially in susceptible populations that have preexisting glaucoma or limited corneal endothelial cell reserves, are not completely understood. Significant retinal nerve fiber layer loss has occurred with large spikes in IOP following IVT injections, and ACPs done before the IVT injection can prevent IOP spikes and retinal nerve fiber loss. ⁴ Patients requiring regular IVT injections to treat chronic retinal diseases may need numerous ACPs over a lifetime, and the effects on corneal endothelial cell counts are unknown. Our study aimed to quantify corneal endothelial cell loss (ECL) following a single needle puncture through the clear cornea. We assessed the average size of the wound area resulting from needle punctures through porcine corneas. We extrapolated these findings to estimate the percentage of ECL from a single needle stick that would occur in a healthy adult human cornea.

Methods

Porcine eyes were purchased from Sierra for Medical Science (Whittier, CA, USA). Nine porcine corneas were used, and each received two paracenteses. The eyes were held with a gauze pad and punctured with a 27-gauge (27G) needle (Monoject^TM, St. Louis, MO, USA) in two corneal quadrants. The needle puncture track was made through the clear cornea parallel to the iris. The needle was retracted immediately after penetration. A 4 mm trephine was used to isolate the needle-punctured areas, and the corneal buttons were placed endothelial surface upwards on a microscope slide. To aid in the visualization of the puncture wounds under a digital light microscope, the endothelial cell surface was dual-stained with Trypan Blue (cell viability stain; Sigma-Aldrich Cat. #T 0776, St. Louis, MO, USA) and Alizarin Red S (intercellular border stain; Sigma-Aldrich Cat. #T 0776, St. Louis, MO, USA), as previously described.^5,6 In brief, a 0.25% stock solution of Trypan Blue was prepared by dissolving 0.25 g of the dye in 100 mL of 0.9% saline. A stock solution of Alizarin Red S (1%) was prepared in 100 mL buffered saline solution (BSS; Alcon Laboratories, Fort Worth, TX, USA) by stirring for several hours with a magnetic stirrer. The solution was filtered (8 µm filter) to remove undissolved dye, and the stock solution was stored at room temperature. A working solution of the Alizarin Red S stain was prepared by titrating 5 mL of stock stain with ammonium hydroxide (0.1% in BSS) to pH 4.2. This pH adjustment was necessary for the optimal staining of the intercellular borders. The endothelium of the corneal tissue was covered with Trypan Blue for 1.5 to 3 minutes. The stain was poured off and rinsed with phosphate buffered saline (PBS). The endothelium was covered with Alizarin Red S for 3 minutes and rinsed thoroughly with PBS. Afterward, the corneas were immediately photographed with a digital light microscope. The puncture area through Descemet’s membrane was clearly visible and stained with Trypan Blue (Fig. 1). Extending beyond the puncture site was a surrounding area of ECL that stained with Alizarin Red, and no distinct cell borders were visible. The images were calibrated, and the wound area, including the surrounding area of ECL around the puncture site, was digitized and measured with Adobe Photoshop CS4 software. The wound areas were quantified in square millimeters (mm²) and then extrapolated to predict ECL in a healthy adult human cornea. Descriptive statistics for the measured wound area were expressed as mean ± standard deviation.

FIG. 1.

Digital image with a light microscope of the posterior aspect of a porcine cornea that has been punctured with a 27G needle. Trypan Blue staining visualizes the cut edges of Descemet’s membrane at the puncture site. The location of the puncture site is labeled with a yellow line. Diffuse, pink-red staining of the damaged area surrounding the puncture site (black arrows) represents the total area of endothelial cell damage from the needle puncture. The inserted measuring scale is 10 μm per division.

The average wound area observed after clear cornea puncture was used to predict ECL in a human cornea, assuming an average central endothelial cell density of 2,500 cells/mm² and 325,000 total corneal endothelial cells. ⁷ Data were further extrapolated to predict the number of cells lost and the percentage of ECL expected from various clinically applicable standard needle gauges. ⁸ For all extrapolations, a linear relationship was assumed between the cross-sectional area of the needle and the average wound area observed with 27G needle punctures.

Results

Nine porcine eyes were used, with two needle sticks per eye (N = 18 needle punctures). Microscopic images demonstrated the tissue trauma to the corneal endothelium created by the needle, as shown in Figure 1. The linear cut observed in Descemet’s membrane approximately matched the observed outer diameter of the needle (ie, 0.4 mm). There was collateral damage around the linear cut, presumably from the surrounding forces around the cut area as the needle passed through Descemet’s membrane. Diffuse and pink-red staining of the damaged area surrounding the cut was consistent with exposed Descemet’s membrane due to the detachment of endothelial cells (ie, denuded endothelium). ⁹ The reported wound area was 0.274 ± 0.122 mm². Extrapolating to humans with an average of 2,500 cells/mm², 685 cells were lost, representing 0.21% of the total endothelium, assuming an average of 325,000 cells per cornea.

The observed area of ECL was larger than the cross-sectional area of the 27G needle (0.12 mm²). The cross-sectional areas of the standard 28G, 30G, and 33G needles are 0.102 mm², 0.071 mm², and 0.031 mm², respectively. Extrapolating from the damaged area observed with corneal needle puncture using a 27G needle, the predicted wound areas for the 28G, 30G, and 33G needles were 0.234 mm², 0.163 mm², and 0.072 mm², respectively. The predicted cell loss for 28G, 30G, and 33G needle stick corresponded to a loss of 586 cells (0.18% of the corneal endothelium), 407 cells (0.13% of the corneal endothelium), and 181 cells (0.06% of the corneal endothelium), respectively.

Discussion

The preferred volume for injecting drugs into the vitreous is 0.05 mL or less. Volumes greater than this increase the risk of IOP spikes that can lead to eye pain and transient central artery occlusion with temporary vision loss. ¹⁰ Volumes of 0.1 mL or more injected into the vitreous of human eyes, such as those used with the currently approved anti-complement therapies, have a greater propensity to cause significant IOP spikes. ¹¹ In a study using triamcinolone acetonide, IOP measurements were performed after an IVT injection of 0.1 mL and ranged from 35.2 to 89.0 mmHg. ¹² Eyes with shorter axial lengths, such as those with hyperopes, have a greater propensity to IOP spikes ¹² due to a lower vitreous volume and higher scleral rigidity. ¹³ ACP can be used before an IVT injection to prevent IOP spikes in at-risk patients or after an IVT injection if a patient has sustained vision loss from central artery occlusion due to high IOP.^4,13

ACP is effective at preventing or treating IOP spikes following IVT injections, but this comes with the liability of ECL with each puncture. The corneal endothelium is critical in maintaining corneal transparency and preserving visual acuity. The human corneal endothelium undergoes an annual decline in cell density of approximately 0.3% to 0.6% and does not have regenerative abilities. ¹⁴ This poses a significant challenge, as endothelial dystrophies, ocular and systemic diseases, and accidental or surgical trauma can further accelerate ECL, ultimately compromising corneal function and vision.^15–17 The long-term effects of ACP on the corneal endothelial cell counts associated with IVT injections have not been studied, but there is evidence of common treatment patterns. One study reported that patients averaged 45 injections during treatment. ¹⁸ Furthermore, a 10-year study on anti-vascular endothelial growth factor (VEGF) treatment revealed that patients could receive an average of 52.2 injections over the decade. ¹⁹ These findings highlight the extensive and ongoing nature of the treatment required to manage neovascular age-related macular degeneration effectively. If a patient requires an ACP following ∼50 anti-VEGF injections in one eye over time with a 27G needle, there is an ECL of 0.21% per needle stick, and the patient will lose 10.5% of cells per cornea. Frequent ACPs following IVT injections may be associated with clinically significant ECL; thus, caution is advised, particularly in patients with compromised corneas and low endothelial cell counts.

The limitations of this study include the extrapolation of data from a porcine model to predict ECL in humans. Nevertheless, porcine corneas are similar to human corneas in terms of size, structure, function, and corneal endothelial cell density. ²⁰ This similarity and availability make porcine eyes a valuable model for ophthalmological research and model development. Another limitation is the assumption that the size of the damaged area in a porcine cornea with a needle puncture would be similar in humans with a limited corneal endothelial cell reserve, for example, in a patient with Fuch’s endothelial corneal dystrophy. The methods and results of this porcine model could be helpful in potentially examining human eyes from human eye banks with diseased corneas to determine if there is greater ECL with needle sticks associated with ACP. This study did not directly measure IOP when passing the needle through the cornea, and differences in IOP may have had an impact on the area of ECL. Finally, a limitation of this study is the variability observed in the average wound area of puncture sites, where the standard deviation amounted to half the mean wound size after needle puncture. Biological measurements frequently display moderate to substantial variability due to factors such as individual differences, genetic variation, environmental conditions, and other experimental influences. Because the number of corneal samples in this study was limited, increasing the sample size would likely yield a more precise estimate of the mean wound size.

Conclusions

This study reports the first known extrapolation of ECL from needle sticks with an ACP procedure using a porcine model. Frequent IVT injections are required for anti-VEGF and anti-complement drugs to treat chronic retinal diseases, and eyes at risk of complications related to high IOP spikes, such as hyperopic eyes or those with advanced glaucoma, may need frequent ACPs. In patients receiving frequent ACPs, clinically significant ECL may occur, and caution is advised, particularly in patients with compromised corneas and low endothelial cell counts. Future studies utilizing human cadaveric corneas obtained from eye banks may help validate our findings and highlight the pathway toward clinical translation.

Authors’ Contributions

E.B.R.: Writing—original draft, writing—review and editing, visualization. A.N.: Writing—review and editing, supervision. M.R.R.: Conceptualization, methodology, formal analysis, writing—review and editing, supervision.