Incidence of Incision-Related Descemet Membrane Detachment Using Phacoemulsification With Trapezoid vs Conventional 2.2-mm Clear Corneal Incision: A Randomized Clinical Trial

Key Points

Question  Does phacoemulsification with trapezoid incision reduce the risk of incision-related Descemet membrane detachment (DMD) effectively and safely compared with conventional 2.2-mm clear corneal incision?

Findings  This randomized clinical trial including 130 eyes from 130 patients found phacoemulsification with a trapezoid incision shape significantly reduced the incidence of incision-related DMD at postoperative day 1 without increasing other intraoperative and postoperative complications, such as unstable anterior chamber, incision leakage, posterior capsular rupture, surgically induced astigmatism, or central corneal endothelial cell loss.

Meaning  These results suggest phacoemulsification with a trapezoid incision reduces the incidence of incision-related DMD effectively at postoperative day 1; the clinical relevance remains to be determined.

Importance  The conventional 2.2-mm clear corneal incision is relatively narrow compared with the sleeves of Phaco handpieces, resulting in friction at the incision site and increased risk of incision-related Descemet membrane detachment (DMD). The modified 2.2-mm incision only enlarged internal width to 3.0 mm, forming a trapezoid incision shape, which may reduce the friction of surgical instruments and decrease the risk of incisional DMD.

Objective  To compare the incidence of incision-related DMD between eyes undergoing modified vs conventional 2.2-mm incision phacoemulsification for hard nuclear age-related cataract.

Design, Setting, and Participants  This double-masked, parallel randomized clinical trial was conducted from July 22, 2019, to January 22, 2020, at Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, China. The study included patients with age-related cataract and nuclear opalescence grade of 4.0 or greater based on the Lens Opacities Classification System III. Patients were enrolled in this study according to the following inclusion criteria: (1) age between 65 to 90 years; (2) pupil size of 6 mm or greater after dilation; (3) Lens Opacities Classification System III nuclear opalescence grade of 4.0 or more; and (4) corneal endothelial cell density greater than 1500 cells/mm².

Interventions  Modified (enlarged internal width to 3.0 mm) or conventional 2.2-mm incision phacoemulsification with intraocular lens implantation.

Main Outcomes and Measures  Incidence of incision-related DMD at postoperative day 1.

Results  A total of 130 eyes of 130 patients were randomized into the conventional group (n = 65) or the modified group (n = 65). The mean (SD) age of participants was 74.5 (5.9) years and 74.3 (6.0) years in the conventional and modified groups, respectively. A total of 26 participants in the conventional group (40%) and 27 in the modified group (42%) were men. Compared with eyes in the conventional group, the incidence of DMD in eyes in the modified group was significantly lower at postoperative day 1 (difference, 26.15; 95% CI, 9.60-42.71; P = .003). The difference at postoperative day 7 was 16.92 (95% CI, 2.91-30.94; P = .02). The length of DMD (postoperative day 1: difference, 0.188; 95% CI, 0.075-0.301; P = .002) and maximal corneal thickness at incision site (postoperative day 1: difference, 0.032; 95% CI, 0.006-0.057; P = .02; postoperative day 7: difference, 0.019; 95% CI, 0.003-0.035; P = .02) were lower in the modified group, while visual quality parameter modulation transfer function (postoperative day 1: difference, −0.033; 95% CI, −0.064 to −0.001; P = .04) was higher. No difference was observed between the 2 groups in best-corrected visual acuity, central corneal endothelium loss, or surgically induced astigmatism at any follow-up time. There were no intraoperative complications in the 2 groups.

Conclusions and Relevance  These findings suggest that modified 2.2-mm trapezoid incision phacoemulsification reduces the incidence of DMD for hard nuclear age-related cataract at postoperative day 1 and might be considered in patients at high risk of incision-related DMD, although the clinical relevance cannot be determined with certainty from this trial.

Trial Registration  ClinicalTrials.gov identifier: NCT04014699

Phacoemulsification is the most commonly performed surgery in modern ophthalmology.^1,2 Surgical incision construction is one of the fundamental components of phacoemulsification, and the 2.2-mm clear corneal incision is a common approach in current clinical practice for its slight surgically induced astigmatism (SIA) compared with 3.0-mm and other incision sizes.^3,4 However, the 2.2-mm incision is relatively narrow compared with the sleeves of Phaco handpieces,^3,5 resulting in significant friction at the incision site and increased risk of incision-related Descemet membrane detachment (DMD).⁶ The occurrence of DMD at the incision site, present at higher incidence among inexperienced surgeons and in patients with hard nuclear age-related cataract, can hinder incisional healing.^7-9 Severe DMD can also lead to postsurgical corneal edema, reduced vision, and even corneal decompensation requiring corneal transplant.^10,11 Therefore, the conventional 2.2-mm phacoemulsification incision can still benefit from further modification.

We enlarged the internal incision width to 3.0 mm during phacoemulsification surgery, resulting in a trapezoidal tunnel incision shape. This allows a larger range of motion for surgical instruments, avoiding friction at the incision site, which may reduce the risk of surgical complications, including DMD. We performed a double-masked, parallel randomized clinical trial to investigate whether this enlarged internal incision width could decrease the incidence of incision-related DMD without increasing other surgical complications, such as SIA and central corneal endothelial cell loss, compared with phacoemulsification surgery using the conventional 2.2-mm incision.

This is a double-masked, parallel randomized clinical trial performed at the Zhongshan Ophthalmic Center in Guangzhou, China. The protocol was approved by the Ethics Committee at Zhongshan Ophthalmic Center and can be found in Supplement 1, and written informed consent was obtained from all participants. Participants were not offered any compensation or incentives to participate. All research procedures followed the Declaration of Helsinki. This study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline.

From July 22, 2019, to January 22, 2020, data from patients with age-related cataract scheduled for phacoemulsification and intraocular lens implantation were screened. Only right eyes were included for patients undergoing surgery in both eyes. Patients were enrolled in this study according to the following inclusion criteria: (1) age between 65 to 90 years; (2) pupil size of 6 mm or greater after dilation; (3) Lens Opacities Classification System III (LOCSIII) nuclear opalescence grade of 4.0 or more¹²; and (4) corneal endothelial cell density greater than 1500 cells/mm². Patients with a history of ophthalmic trauma or surgery; other ocular diseases, such as glaucoma, uveitis, or high myopia; and other ocular factors that would make surgery challenging or dangerous, including small pupil and shallow anterior chamber, were excluded (Figure 1).

The sample size was calculated for our primary outcome on the basis of previous reports stating incidence of incision-related DMD at postoperative day 1 was near 50%.⁴ Assuming incidence of incisional DMD at 1 day postoperatively was 50% in the conventional group and 20% in the modified group, a 2-sided significance level of .05, power of 90%, and 20% loss to follow up, we calculated that a sample size of 65 participants would be needed in each group.

An online random number generator¹³ was used to assign study participants to either the conventional or modified group in a 1:1 ratio, with permuted block size of 4. Allocation codes were sealed in sequentially numbered opaque envelopes. The participating surgeon was informed of the study group allocation just prior to surgery. The statistician who generated the allocation schedule was not involved in the assignment of treatments to participants. The surgeon and statistician did not participate in postsurgical follow-up examinations. Data were entered into electronic spreadsheets by study personnel who were masked to patient group assignment and were checked by the study coordinator. All participants, caregivers, investigators, and clinical staff remained masked to patient study group assignment until trial completion.

All procedures for study participants were performed by an experienced ophthalmologist (Y.L.) following standardized procedures. All patients were operated on using topical anesthesia with proxymetacaine hydrochloride, 0.5% (Alcaine; Novartis). All surgical incisions were created with a 2.2-mm keratome (Alcon Labs; Clear Cut Intrepid SB) for its common incision size in clinic.³ A temporal 2-plane clear corneal incision was made in eyes assigned to the conventional group with the internal and external incision width both of 2.2 mm; in eyes assigned to the modified group, the internal width of the incision was enlarged to 3.0 mm measured by a ruler, while the external width remained at 2.2 mm, forming a trapezoidal incision shape (eFigure 1 in Supplement 2). The tunnel length in both groups was 1.7 mm. Injection of an ophthalmic viscoelastic device consisting of medical sodium hyaluronate gel (Hangzhou Singclean Medical Products Co) was used to maintain the stability of the anterior chamber. A 26-gauge capsulotomy needle was used to create a continuous circular capsulorhexis of 5.5 to 6.0 mm in diameter. A Centurion Vision System (Alcon Labs) device was used to perform phacoemulsification surgery, including nucleus chopping with a 0.9-mm ultrasonic tip (Centurion OZil handpiece; Alcon Labs) and a straight-headed coaxial tip for irrigation-aspiration with noncontinuous irrigation. Torsional phacoemulsification was set between 60% and 100%, suction velocity was 33 to 35 mL/min, and negative pressure was maintained in the range of 330 to 350 mm Hg during phacoemulsification. A single-focus intraocular lens (SN60WF; Alcon Labs) was implanted. Intraoperative parameters, including total use of balanced salt solution (BSS), cumulative dissipated energy (CDE), ultrasound time (UST), and surgical complications, were recorded. An intraoperative optical coherence tomography system (OCT; Zeiss Rescan 700; Carl Zeiss Meditec) was connected to the operating microscope to observe the main corneal incision during surgery. Both groups were treated with tobramycin and dexamethasone eye drops 4 times per day for 1 week (Alcon) after surgery. Identical ocular examinations were performed on both groups at postoperative day 1, postoperative day 7, postoperative month 1, and postoperative month 3.

The main outcome measure was the incidence of DMD at postoperative day 1 observed by an anterior-segment OCT (AS-OCT) (CASIA1; Tomey). Secondary outcome measures included the length of DMD defined as the maximum cross-sectional length across the whole DMD area based on the height-, extent-, length-, and pupil-based (HELP) algorithm,¹⁴ maximal corneal thickness at the incision site, best-corrected visual acuity (BCVA), SIA, central corneal endothelial cell loss, and modulation transfer function (MTF). The length of DMD and the maximum corneal thickness at the incision site were evaluated by an AS-OCT (CASIA1). Central corneal endothelial cell density was measured by noncontact specular microscopy (SP-1P; Topcon), and central corneal endothelial loss was calculated according to the preoperative and postoperative endothelial cell density. BCVA and uncorrected visual acuity were assessed using an Early Treatment Diabetic Retinopathy Study visual acuity chart. Corneal astigmatism was measured with AS-OCT (CASIA1), and SIA was calculated at each postoperative visit using the following equation: K₂ = (K₁² + K₃² − 2K₁ × K₃cos[2θ₃ − 2θ₁])^0.5.¹⁵ Itrace (Tracey Technologies) was used to measure the visual quality parameter MTF.

Baseline and intraoperative characteristics were described as means and SDs for continuous variables with normal distribution and counts and percentages for categorical variables. The normality of continuous data’s distribution was checked by Shapiro-Wilk normality test and histogram. Analysis of primary and secondary outcomes was performed by the principle of intention to treat, which required data from all participants who underwent randomization to be included in the final analysis. All missing data were imputed by creating 20 copies of the data, in which missing values were imputed using multiple imputation.¹⁶ Different models were used for imputation, including logistic regression for binary data and linear regression for continuous data, selecting the independent variables based on predictive value and availability of data. Final results were obtained by averaging results from each of these 20 data sets using Rubin rules.¹⁶ The group difference in incidence of DMD and 95% CIs were calculated by proportion z test, and when 1 or more of the cell counts in the 2 × 2 table was less than 5, the Exact method was used instead of proportion z test. Histogram and line plot were made to show the group differences. Two-sample t tests were conducted to estimate the mean group difference and 95% CIs for secondary outcomes. Most cases of DMD were self-healing at the end of study in both groups, and the survival time could not be calculated. Therefore, Kaplan-Meier estimation, which had been prespecified in the study protocol, was not used in this study. A 2-sided P value less than .05 was considered to be statistically significant for the primary outcome, with no adjustment for the multiple testing for secondary or exploratory outcomes. Statistical analyses were performed using Stata version 13.0 (StataCorp) and SAS version 9.4 (SAS Institute).

A total of 130 eyes from 130 patients were randomized into the conventional group (n = 65) or the modified group (n = 65). Among these patients, 2 were lost to follow-up at postoperative day 7 owing to lost contact; 4 were lost to follow-up at postoperative month 1, including 1 refusal and 3 owing to lost contact; 6 were lost to follow-up at postoperative month 3, including 4 refusals and 2 owing to lost contact. A total of 12 patients (9.2%) were lost to follow-up, and the remaining 118 participants (90.8%) completed all study visits (Figure 1).

The mean (SD) age of participants was 74.5 (5.9) years and 74.3 (6.0) years in the conventional and modified groups, respectively. A total of 26 participants in the conventional group (40%) and 27 in the modified group (42%) were men. All procedures were successfully completed, and no intraoperative complications occurred in either group (Table 1).

As shown in Figure 2, the incidence of DMD at postoperative day 1 was significantly higher in the conventional group (43 of 65 [66%]) compared with the modified group (26 of 65 [40%]; difference, 26.15; 95% CI, 9.60-42.71; P = .003). The incidence of DMD at postoperative day 7 in the conventional group (20 [31%]) was also higher than that in the modified group (9 [14%]; difference, 16.92; 95% CI, 2.91-30.94; P = .02). DMD resolved gradually with longer follow-up time in both groups. The incidence of DMD was 5% (3 of 65) in conventional group and 3% (2 of 65) in modified group at postoperative month 1 and postoperative month 3 (difference, 1.54; 95% CI, −7.68 to 9.68; P > .99).

eFigure 2 in Supplement 2 depicts measurement of DMD by AS-OCT in 2 typical eyes. The length of DMD in the conventional group was longer than that in the modified group at postoperative day 1 (mean [SD] length, 0.787 [0.222] mm vs 0.599 [0.238] mm; difference, 0.188; 95% CI, 0.075-0.301; P = .002). No difference was detected at subsequent visits (Table 2).

eFigure 2 in Supplement 2 depicts measurement of maximal corneal thickness at the incision site by AS-OCT in 2 typical eyes. The maximal corneal thickness in the conventional group was larger than that in the modified group at postoperative day 1 (mean [SD] size, 0.972 [0.070] mm vs 0.940 [0.079] mm; difference, 0.032; 95% CI, 0.006-0.057; P = .02) and postoperative day 7 (mean [SD] size, 0.838 [0.052] mm vs 0.819 [0.039] mm; difference, 0.019; 95% CI, 0.003-0.035; P = .02). There was no difference in maximal corneal thickness between the 2 groups at subsequent follow-up visits (Table 2).

There was no difference in central corneal endothelial loss between the 2 study groups at any follow-up time (Table 2).

The mean BCVA improved in both groups after surgery, but there were no differences in BCVA or SIA between study groups at any follow-up visit. SIA decreased gradually over follow-up in both groups (Table 2).

MTF was better in the modified compared with the conventional group at postoperative day 1 (difference, −0.033; 95% CI, −0.064 to −0.001; P = .04). No difference was observed at subsequent visits (Table 2).

Results of our study demonstrated that, compared with conventional 2.2-mm clear corneal incision, modified trapezoidal 2.2-mm incision could effectively reduce the incidence of DMD at postoperative day 1 after phacoemulsification surgery among patients with hard nuclear cataract, although the clinical relevance remains to be determined. There was no adjustment to P values for multiple comparisons for the secondary outcomes and this study could not determine if the differences at postoperative day 1 definitively remained thereafter, and therefore, the clinical relevance of these results cannot be determined with certainty.

Enlarging the internal incision width to 3.0 mm to form a trapezoidal tunnel can increase the range of motion for surgical instruments and reduce mechanical injury from compression of the incision. SIA is largely determined by the width of the external incision wound at the anterior corneal surface.^17,18 Therefore, we proposed to expand the internal incision width of the conventional 2.2-mm clear corneal tunnel incision, hypothesizing that the impact on SIA would be minimal.

DMD is an important indicator of incision-related complications in phacoemulsification.^4,19,20 Severe incision-related DMD will lead to corneal endothelial avulsion (Video) and damage corneal endothelial cells.¹⁹ Previous studies found that the highest incidence of DMD after phacoemulsification occurred on postoperative day 1.^4,21,22 Therefore, the incidence of DMD at postoperative day 1 was selected as the primary outcome measure in this study. Our previous work and that of others have found that increasing age, harder nuclear cataract, and higher levels of intraoperative ultrasound energy are independent risk factors for DMD.^4,21,23,24 Therefore, patients with hard nuclear cataract were enrolled in this study. We estimated a DMD rate of 50% in the conventional group based on the previous study in which the participants were patients with cataract with a mean LOCSIII nuclear opalescence grade of 3.0. However, as mentioned above, our study included patients with cataract with a nuclear opalescence grade of 4.0 or greater.⁴ Therefore, our study rate was higher than anticipated.

The incidence of DMD was significantly lower in the early postoperative period in the modified group, which we speculate could be because greater internal incision width increased the range of motion for surgical instruments and reduced mechanical injury–caused wound compression during the surgery or because the larger internal incision increased flow of BSS through the incision tunnel, potentiating the absorption of thermal energy from the ultrasound and reducing damage to the incision caused by local temperature increase. We did not find any difference in ultrasonic energy, BSS volume, or UST between the 2 groups, potentially calling into question the hypothesis related to increased BSS flow. The volume of liquid irrigated into as well as drained and leaked out of the eye in both groups was not measured. We will measure this outcome in our next study. Additionally, we found that intraoperative occurrence of DMD was observed later in the modified group compared with the conventional group (eFigure 3 in Supplement 2). The length of DMD at the end of surgery was shorter in the modified group (eFigure 3 in Supplement 2). These findings were consistent with the hypothesis that the modified incision was beneficial in preventing DMD. This study included patients with cataract with healthy corneas; therefore, the DMD would be virtually gone in the short term. However, the rate and extent of DMD would be much more serious and could even lead to corneal decompensation for patients with unhealthy corneas and those treated by inexperienced surgeons if the instruments enter and exit the incision repeatedly during the procedure.⁶

Despite many previous studies describing the dynamic characteristics of various clear corneal incision sizes, few reports have focused on stability of 2.2-mm corneal incisions.^24,25 In this study, we used high-resolution AS-OCT to assess details of corneal structure, including DMD length and maximum corneal thickness at the incision site. The incision-related DMD was floating in the aqueous humor, which led to the difficulty of measurement of DMD area. Therefore, we used the length of DMD as an indicator of the extent of DMD based on the HELP algorithm.¹⁴ It has been reported in our previous studies and those of others that at postoperative day 1, the incidence of DMD with 2.2-mm incisions is higher than with 2.85-mm incisions, and maximum corneal thickness at the site of 2.2-mm incisions is lower than for 1.8-mm incisions.^3,4 These findings suggest smaller incision width can lead to more severe incisional injury and poorer incision architecture stability.^3,4 This is consistent with the results of the current study.

Previous studies have also suggested that incision size is closely related to increased risk of corneal endothelial loss.²⁶ Lee and colleagues²⁷ reported that for hard nuclear cataract, phacoemulsification with a 1.8-mm incision required longer UST and higher mean CDE compared with 2.2-mm incisions, leading to greater corneal endothelial cell loss. However, our previous study found that there was no difference in UST, CDE, or corneal endothelial loss for phacoemulsification with 3.0-mm, 2.2-mm, and 1.8 mm incision.³ This study found that there was no difference in these surgical parameters and corneal endothelial loss between the modified and conventional group, further confirming our previous findings and also affirming the safety of the modified 2.2-mm incision. The inconsistency between our findings and those of other studies may be owing to difference in Phaco systems and surgeons’ experiences; our conclusions require verification in larger cohorts.

We found that the modified 2.2-mm incision resulted in higher MTF, which reflected less high-order aberrations compared with the conventional 2.2-mm incision at postoperative day 1, suggesting that the modified incision improved the early postoperative visual quality of patients. This may be attributed to reduced incisional injury and impact on corneal morphology in the early postoperative period. Given that corneal edema and DMD are commonly seen in conventional 2.2-mm Phaco incisions, minimizing the risk of such complications is of great clinical importance for postoperative visual recovery. However, we found that there was no difference in SIA or BCVA between the modified and conventional group across the follow-up period, suggesting that this slight change of the 2.2-mm incision architecture in the modified group does not affect visual recovery of patients.

Strengths of this study include the homogeneity of participant characteristics between groups; surgical procedures, postoperative medical regimen, and data analysis were also standardized. Limitations of this study must also be acknowledged. First, all procedures were performed by 1 experienced ophthalmologist to ensure the uniformity of interventions in this study. Incision-related DMD could easily be ignored by experienced surgeons owing to the mild degree, but for inexperienced surgeons, the DMDs usually are severe owing to the unskilled operations. Therefore, our study results need to be further verified among surgeons with different levels of experiences. Second, this study only included patients with age-related cataract with hard nuclei and normal endothelial cell density without any other corneal diseases. Results need to be verified in patients with unhealthy corneas with higher risk of DMD.

In conclusion, we demonstrated that compared with conventional 2.2-mm clear corneal incision, phacoemulsification with modified trapezoidal 2.2-mm incision reduced the incidence of incision-related DMD effectively at postoperative day 1. Although the clinical relevance of these results cannot be determined with certainty, no safety problems were identified.

Accepted for Publication: August 29, 2021.

Published Online: October 14, 2021. doi:10.1001/jamaophthalmol.2021.4148

Corresponding Author: Yizhi Liu, MD, PhD (yzliu62@yahoo.com), and Lixia Luo, MD, PhD (luolixia@gzzoc.com), Zhongshan Ophthalmic Center, Sun Yat-Sen University, No. 7 Jinsui Rd, Guangzhou 510000, China.

Author Contributions: Drs Luo and Y. Liu had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Dai, Z. Liu, and W. Wang contributed equally.

Study concept and design: Dai, Z. Liu, W. Wang, L. Wang, Zhang, Congdon, He, Luo, Y. Liu.

Acquisition, analysis, or interpretation of data: Dai, Z. Liu, W. Wang, Han, L. Jin, Chen, G. Jin, Qu, J. Liu, Y. Liu.

Drafting of the manuscript: Dai, Z. Liu, W. Wang, Chen, G. Jin, L. Wang, Zhang, Y. Liu.

Critical revision of the manuscript for important intellectual content: Dai, Z. Liu, W. Wang, Han, L. Jin, L. Wang, Qu, J. Liu, Congdon, He, Luo, Y. Liu.

Statistical analysis: Dai, W. Wang, Han, L. Jin, G. Jin, L. Wang, Zhang, Qu, Congdon.

Obtained funding: Dai, Z. Liu, Y. Liu.

Administrative, technical, or material support: W. Wang, Chen, G. Jin, Qu, J. Liu, Y. Liu.

Study supervision: He, Luo, Y. Liu.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by the Construction Project of High-Level Hospitals in Guangdong Province (grant 303020102) and the National Natural Science Foundation of China (grants 81900815 and 81873675). Dr Congdon is supported by the Ulverscroft Foundation.

Role of the Funder/Sponsor: The Construction Project of High-Level Hospitals in Guangdong Province had a role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. The National Natural Science Foundation of China and Ulverscroft Foundation had a role in the collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 3.