Original Article

Taiwan J Ophthalmol 2024;14:83‑87

Access this article online

Quick Response Code:
Risk factors for corneal endothelial cell
loss after phacoemulsification
Natalie Si‑Yi Lee1,2, Keith Ong3,4,5,6*

Website: Abstract:
http://journals.lww.com/TJOP PURPOSE: The purpose of this study was to evaluate the changes in corneal endothelial cell
DOI:
density (CECD) occurring after cataract phacoemulsification surgery and identify factors associated
10.4103/tjo.TJO-D-23-00146 with cell loss.
MATERIALS AND METHODS: This was a retrospective study involving patients who underwent
cataract phacoemulsification surgery between January 1, 2018, and December 31, 2018, at two private
hospitals. Demographic data and biometric parameters were obtained preoperatively. Ultrasound
metrics were recorded for each operation, including total on time (TOT), total equivalent power in
position 3, and cumulative dissipated energy (CDE). Using corneal specular microscopy, CECD
was measured preoperatively and postoperatively at 12, 24, and 36 months. Factors associated
with decreased CECD were identified.
RESULTS: This study included 223 eyes of 133 patients. The mean CECD was 2530.03 ± 285.42 cells/mm2
preoperatively and significantly decreased to 2364.22 ± 386.98 cells/mm2 at 12 months (P < 0.001),
2292.32 ± 319.72 cells/mm2 at 24 months (P < 0.001), and 2242.85 ± 363.65 cells/mm2 at
36 months (P < 0.001). The amount of cell loss was associated with age, gender, preoperative
CECD, preoperative anterior chamber depth, lens thickness, TOT, and CDE. Using multivariate
analysis, age, preoperative CECD, and TOT were identified as independent predictors for CECD
loss 12 months after surgery.
CONCLUSION: The greatest decrease in CECD occurred during the first year after cataract surgery,
and the amount of cell loss was influenced by both baseline patient characteristics and ultrasound
metrics. Longer‑term prospective studies in a larger cohort may yield more information.
Keywords:
Cataract surgery, corneal endothelium, phacoemulsification
1
Faculty of Medicine,
University of New South
Wales, 2Department of Introduction disrupted, the remaining cells enlarge to
Ophthalmology, Royal compensate for missing cells, and there is

C
North Shore Hospital, ataract surgery, the most common an overall decrease in corneal endothelial
3
Northern Clinical
School, University of elective procedure performed cell density (CECD).[2]
Sydney, 4Department worldwide, has the potential to greatly
of Ophthalmology, improve visual function and treat blindness. Normal CECD is approximately
Chatswood Private However, as with any surgical procedure, 2500 cells/mm2, with cell density naturally
Hospital, 5Department of
Ophthalmology, Sydney and despite modern phacoemulsification declining as part of the aging process. [2]
Adventist Hospital, 6Save techniques, postoperative complications However, cataract surgery has been reported
Sight Institute, University may occur.[1] The corneal endothelium is to accelerate this process, with some studies
of Sydney, Sydney, particularly vulnerable to injury during recording rates of cell loss >20% within
Australia
cataract surgery and does not have any the first 12 months postoperatively. [1,3]
*Address for regenerative ability. As part of this process, Damage to corneal tissue may occur due
correspondence: the normal hexagonal cell arrangement is to a combination of factors, including
A/Prof. Keith Ong, direct instrumental trauma, exposure to
2 Railway Avenue, ultrasound energy, free radical formation,
Eastwood, NSW 2122, This is an open access journal, and articles are
Australia. and contact with the intraocular lens or
distributed under the terms of the Creative Commons
E‑mail: drkeithong@ Attribution‑NonCommercial‑ShareAlike 4.0 License, which nuclear fragments.[4,5] When cell density
optusnet.com.au allows others to remix, tweak, and build upon the work
non‑commercially, as long as appropriate credit is given and How to cite this article: Lee NS, Ong K. Risk factors for
Submission: 21‑09‑2023
the new creations are licensed under the identical terms. corneal endothelial cell loss after phacoemulsification.
Accepted: 22‑11‑2023
Taiwan J Ophthalmol 2024;14:83-7.
Published: 30-01-2024 For reprints contact: [email protected]

© 2024 Taiwan J Ophthalmol | Published by Wolters Kluwer - Medknow 83

falls below 500 cells/mm2, there is a high risk of corneal 12, 24 and 36 months. CECD was analyzed using an
edema and decompensation.[2,6] automated noncontact specular microscope (Tomey
EM‑3000, Nagoya, Japan) and recorded as the number of
At present, there is much variation in factors that have cells/mm2. Of the 15 photographs taken by the specular
been shown to influence CECD loss after cataract surgery. microscope, only the cell density from the clearest image
Patient‑related factors, including age,[1,2,7] preoperative was used. All measurements and assessments were
anterior chamber depth (ACD),[8] preoperative CECD,[1] conducted by the same ophthalmic surgeon.
axial length,[9] and cataract grade,[2,7] and surgery‑related
factors, including phacoemulsification energy[1,10] and Statistical analysis
phacoemulsification time,[11,12] have been implicated Demographic data were analyzed using descriptive
in the current literature. However, findings have been statistics. The paired t‑test was used to assess differences
limited by relatively short follow‑up, with studies in CECD preoperatively and at 12, 24, and 36 months
generally reporting CECD outcomes at 12 months or after surgery.
less.[2,4,9,10] Furthermore, with the development of modern
surgical techniques and technologies, some earlier Differences in corneal endothelial cell loss between
studies no longer reflect the current practice.[7,10] groups were evaluated using the independent t‑test.
Pearson’s correlation coefficient was used to identify
The aims of this study were to evaluate the 3‑year significant associations between the amount of CECD loss
changes in CECD after cataract phacoemulsification at 12, 24, and 36 months, and baseline ocular parameters
surgery in a contemporary cohort and identify factors and ultrasound metrics. Variables with P < 0.20 on
affecting cell loss. Understanding these factors may better univariate analysis at 12 months were entered into
inform approaches to minimizing corneal endothelial multiple regression analysis. Backward variable selection
injury intraoperatively and assist in the identification of was performed to determine the variables independently
patients at greater risk of future complications. associated with CECD loss. The sample size was not
formally calculated as this was a retrospective cohort
Materials and Methods study that utilized all available data.[13]

Study population P < 0.05 was used to indicate statistical significance.

This retrospective observational study included All data were analyzed using IBM SPSS Statistics,
patients who underwent cataract phacoemulsification version 26.0 (IBM Corp, Armonk, NY, USA).
surgery, by a single ophthalmic surgeon, between
January 1, 2018, and December 31, 2018. Cataract Ethics statement
phacoemulsification was performed using the Alcon This study was conducted in accordance with the
Constellation OZil IP system with a 0.9 mm mini‑flared Declaration of Helsinki and was approved by the
45° Kelman® phaco tip through a 2.75 mm temporal Northern Sydney Local Health District Human Research
corneal incision. All patients received Alcon SN60WF Ethics Committee (2021/PID01969). Informed consent
IOL. All patients attended follow‑up during the 3‑year was obtained from all participants.
postoperative period. Exclusion criteria were patients
who had cataract surgery combined with another Results
procedure or those where a superior corneal incision
was used. Patients who received toric lenses or other A total of 223 eyes of 133 patients were included in this
lenses were also excluded. study. Baseline characteristics of the study population
are shown in Table 1. During the 3‑year postoperative
Data collection period, the number of eyes analyzed was 186 eyes
Demographic data were collected on patient gender and at 12 months, 139 eyes at 24 months, and 107 eyes at
age. All patients underwent comprehensive ophthalmic 36 months.
examination 1 week before surgery, where the ocular
parameters ACD, axial length (AL), and central corneal The mean CECD was 2530.03 ± 285.42 cells/mm 2
thickness (CCT) were measured using IOLMaster preoperatively. Significant cell loss occurred over each
700 (Carl Zeiss, Germany). Ultrasound metrics were 12‑month period, with mean CECD decreasing to
obtained from the operative record and included total on 2364.22 ± 386.98 cells/mm2 at 12 months (P < 0.001, paired
time (TOT), total equivalent power in position 3 (TEPP3), t‑test), 2292.32 ± 319.72 cells/mm2 at 24 months (P < 0.001,
and cumulative dissipated energy (CDE). paired t‑test), and 2242.85 ± 363.65 cells/mm 2 at
36 months (P < 0.001, paired t‑test). This represented
Corneal specular microscopy was performed on the the rates of cell loss of 9.46%, 1.82%, and 2.93% during
central cornea preoperatively and postoperatively at each subsequent 12‑month period after surgery. The
84 Taiwan J Ophthalmol - Volume 14, Issue 1, January-March 2024
changes in CECD during the 3 years after surgery are Using univariate analysis, factors significantly associated
shown in Figure 1. with CECD loss were identified at 24 months and
36 months postoperatively. The factors influencing
By univariate analysis, factors demonstrating significant cell loss at 24 months were patient age (P < 0.001),
correlations with decreased CECD at 12 months were gender (P < 0.001), preoperative ACD (P = 0.007),
patient age, gender, preoperative CECD, ACD and lens TOT (P = 0.002), and CDE (P = 0.008). At 36 months, only
thickness (LT), TOT, and CDE. Factors that were not patient gender was found to be significantly associated
found to be significant on univariate analysis were TEPP3, with CECD loss (P = 0.041).
preoperative CCT, and preoperative AL. Using multiple
regression analysis, the variables independently associated Discussion
with CECD loss were age (P < 0.001), preoperative
CECD (P < 0.001), and TOT (P = 0.018). These results are Corneal endothelial cell loss occurring after cataract
summarized in Table 2. The final model (P < 0.001) explained surgery is well documented.[2,7] However, to the authors’
21% of the variance in CECD loss at 12 months (r2 = 0.21). knowledge, the rates of CECD decline over a 3‑year
timeline have not yet been evaluated, and the current
Correlations between mean CECD loss and age, understanding of the patient and surgery‑related factors
preoperative CECD, and TOT are shown in Figure 2. influencing cell loss remains limited.
Preoperative CECD was significantly correlated with
the mean CECD at 12 months after surgery (P < 0.001). The amount of reported cell loss occurring after cataract
phacoemulsification surgery is highly varied, ranging
Table 1: Study population characteristics from 11.6% at 1 month,[12] 7.3% at 12 months,[10] 15.0%
Characteristic N=223 at 6 years,[14] and 20.6% after 10 years.[15] The results of
Number of eyes/patients 223/133 our study are comparable to other records of decreased
Age (years), mean±SD (range) 67.96±7.17 (37–90) CECD at 12 months postoperatively.[2,9] However, while
Gender, n (%) the amount of cell loss was the greatest within the first
Male 97 (43.5) year, cell density continued to significantly decline, at
Female 126 (56.5) 1.82% during the second and 2.93% during the third year
Eye laterality, n (%)
after surgery. As the rate of normal CECD decline has
Right 108 (48.4)
been approximated at 0.5%–1% per year,[8] our findings
Left 115 (51.6)
suggest that accelerated corneal endothelial loss after
CECD (cells/mm2), mean±SD 2530.03±285.42
ACD (mm), mean±SD 3.11±0.40
cataract surgery may persist in the long term. This is
LT (mm), mean±SD 4.60±0.39
similar to the adjusted yearly cell loss rates of 2.06%[15]
CCT (µm), mean±SD 544.27±35.36 and 2.5%[16] published in two 10‑year studies. While the
AL (mm), mean±SD 24.41±1.47 precise mechanisms behind a continued and accelerated
TOT (s), mean±SD (range) 77.44±39.71 (2–331) CECD decline remain unclear, it has been hypothesized
TEPP3 (%), mean±SD (range) 16.40±3.36 (10.5–37.7) that chronic cell loss may be perpetuated by subclinical
CDE, mean±SD (range) 12.92±8.39 (0.93– inflammation,[15] exposure to vitreous, and alterations
63.84) in aqueous humor that compromise endothelial cell
SD=Standard deviation, CECD=Corneal endothelial cell density, nutrition.[16]
ACD=Anterior chamber depth, LT=Lens thickness, CCT=Central corneal
thickness, AL=Axial length, TOT=Total on time, TEPP3=Total equivalent
power in position 3, CDE=Cumulative dissipated energy Of the 10 variables included in the initial univariate
analysis, seven were found to demonstrate significant
Table 2: Multiple regression analysis of factors
affecting corneal endothelial cell density loss
Variable B coefficient (95% CI) SE P
Age 9.87 (5.70–14.06) 2.11 <0.001
Gender
Female Reference ‑
Male 36.83 (−21.38–95.05) 29.50 0.213
Preoperative CECD 0.18 (0.08–0.27) 0.05 <0.001
Preoperative ACD −51.35 (−118.42–15.714) 33.99 0.133
Preoperative LT 11.05 (−62.25–84.36) 37.15 0.766
TOT 0.91 (0.16–1.67) 0.38 0.018
CDE −1.36 (−9.47–6.75) 4.11 0.741
CI=Confidence interval, SE=Standard error, CECD=Corneal endothelial cell
density, ACD=Anterior chamber depth, LT=Lens thickness, TOT=Total on Figure 1: Mean corneal endothelial cell density preoperatively and after surgery. Error
time, CDE=Cumulative dissipated energy bars represent standard errors. CECD = Corneal endothelial cell density

Taiwan J Ophthalmol - Volume 14, Issue 1, January-March 2024 85

 a b

c
Figure 2: Scatterplots showing mean decreases in CECD against age (a), preoperative CECD (b), and total on time (c). CECD = Corneal endothelial cell density

associations with cell loss at 12 months. Preoperative The final multivariate model contained patient age,
ACD was first described by Hwang et al.[8] to influence preoperative CECD, and TOT as significant factors
CECD, where similarly to the present study, eyes affecting cell loss. Age has previously been shown
with shallower anterior chambers exhibited higher to influence postoperative CECD[1,2,19] and has been
amounts of cell loss. This is thought to be due to correlated with higher phacoemulsification energy and
phacoemulsification being conducted in closer proximity cataract grade.[1] This is an important consideration
to the corneal endothelium, increasing the risk of injury when planning the timing of surgery, as there may be
from ultrasound and heat energy, instrumental trauma, an increased risk of corneal complications in patients of
and lens fragments.[8] However, the overall evidence advanced age.[19]
on the impact of ACD is unclear, as ACD was not an
independent predictor on multivariate analysis, and Ultrasound time has also been associated with cell
other studies have not shown any significant association loss in some studies[3,9,10,12] and may indirectly reflect
between ACD and cell loss.[9,17] mechanical injury to corneal endothelium. A higher TOT
may be associated with factors such as increasing age,
Similarly, CDE, patient gender, and LT were correlated LT, and cataract grade.[9] Studies have also demonstrated
with corneal endothelial loss at 12 months on univariate increased TOT when phacoemulsification is performed
analysis but were not retained in the final model. Higher by junior surgeons[12] or in the eyes with small pupils.
CDE has been correlated with greater corneal endothelial However, in this study, all procedures were performed
loss in one study,[3] as has higher ultrasound power.[10,12] This by the same experienced surgeon, and pupils were
may be due to increased heat generation and turbulence adequately dilated.
within the anterior chamber during phacoemulsification.[3]
Our results also identified a positive correlation between Preoperative CECD was recently identified as a novel
LT and cell loss, which has not previously been described, predictive factor for cell loss,[1] despite earlier studies
although increasing LT and lens nucleus density have demonstrating no significant association.[2,12,15,16] In the
been associated with advancing age and shallow anterior study by Chen et al.,[1] the eyes with a preoperative
chambers.[9] In this study, while mean cell loss was CECD between 1000 and 2000 cells/mm 2 exhibited
observed to be higher in males than females, patient gender the greatest decrease in cell density. In the current
has consistently been shown not to independently affect study, at 12 months postoperatively, cell density
CECD in the existing literature.[7,15,16,18] remained significantly and positively correlated with
86 Taiwan J Ophthalmol - Volume 14, Issue 1, January-March 2024
preoperative CECD, despite a small positive correlation in patients with mildly low endothelial cell density. J Clin Med
between preoperative CECD and mean cell loss. This 2021;10:2270.
2. Bourne RR, Minassian DC, Dart JK, Rosen P, Kaushal S,
suggests that where possible, it may be beneficial to Wingate N. Effect of cataract surgery on the corneal endothelium:
plan for earlier cataract phacoemulsification surgery, Modern phacoemulsification compared with extracapsular
as eyes will have a higher baseline CECD. Further cataract surgery. Ophthalmology 2004;111:679‑85.
studies to characterize this association would be 3. Mahdy MA, Eid MZ, Mohammed MA, Hafez A, Bhatia J.
beneficial. Relationship between endothelial cell loss and microcoaxial
phacoemulsification parameters in noncomplicated cataract
surgery. Clin Ophthalmol 2012;6:503‑10.
Strengths of the present study include the use of 4. Dzhaber D, Mustafa O, Alsaleh F, Mihailovic A, Daoud YJ.
routinely collected data and regular follow‑up intervals Comparison of changes in corneal endothelial cell density and
in a contemporary cohort. All surgeries, preoperative central corneal thickness between conventional and femtosecond
measurements, and follow‑up were conducted by laser‑assisted cataract surgery: A randomised, controlled clinical
trial. Br J Ophthalmol 2020;104:225‑9.
the same ophthalmic surgeon, eliminating the impact 5. Takahashi H. Corneal endothelium and phacoemulsification.
of surgeon experience as a confounding factor. [9,12] Cornea 2016;35:3‑7.
Limitations of the present study include its retrospective 6. Lee NS, Ong RM, Ong K. Changes in corneal endothelial cell
nature. Other factors described to influence postoperative density after trabeculectomy. Eur J Ophthalmol 2023;33:2222‑7.
CECD, such as cataract grade,[1,2,7,12] infusion volume,[3,7] 7. Hayashi K, Hayashi H, Nakao F, Hayashi F. Risk factors for
corneal endothelial injury during phacoemulsification. J Cataract
lens size,[7] and diabetes mellitus,[18,19] were unable to be
Refract Surg 1996;22:1079‑84.
evaluated, as were other markers of corneal endothelial 8. Hwang HB, Lyu B, Yim HB, Lee NY. Endothelial cell loss after
integrity, such as cell size and morphology.[5,14] Prospective phacoemulsification according to different anterior chamber
studies examining these variables and outcomes would depths. J Ophthalmol 2015;2015:210716.
provide more comprehensive information and improve 9. Walkow T, Anders N, Klebe S. Endothelial cell loss after
phacoemulsification: Relation to preoperative and intraoperative
predictability. Assessing the impact of longer ultrasound
parameters. J Cataract Refract Surg 2000;26:727‑32.
time and less power, compared to shorter ultrasound 10. Dick HB, Kohnen T, Jacobi FK, Jacobi KW. Long‑term endothelial
time, and more power is also an area of future research, cell loss following phacoemulsification through a temporal clear
as this could not be differentiated in the current study. corneal incision. J Cataract Refract Surg 1996;22:63‑71.
11. Haddad JS, Borges C, Daher ND, Mine A, Salomão M,
Ambrósio R Jr. Correlations of immediate corneal tomography
Conclusion changes with preoperative and the elapsed phaco parameters.
Clin Ophthalmol 2022;16:2421‑8.
Significant corneal endothelial cell loss occurred 12. O’Brien PD, Fitzpatrick P, Kilmartin DJ, Beatty S. Risk factors for
over 3 years after cataract phacoemulsification surgery. endothelial cell loss after phacoemulsification surgery by a junior
The amount of cell loss increased with older age, resident. J Cataract Refract Surg 2004;30:839‑43.
preoperative CECD, and TOT. Planning for earlier 13. Moons KG, Altman DG, Reitsma JB, Ioannidis JP, Macaskill P,
Steyerberg EW, et al. Transparent reporting of a multivariable
surgical intervention and long‑term monitoring of CECD prediction model for individual prognosis or diagnosis (TRIPOD):
may improve patient outcomes. Explanation and elaboration. Ann Intern Med 2015;162:W1‑73.
14. Lundberg B, Behndig A. Intracameral mydriatics in
Data availability statement phacoemulsification cataract surgery – A 6‑year follow‑up. Acta
Ophthalmol 2013;91:243‑6.
The data that support the findings of this study are not
15. Choi JY, Han YK. Long‑term (≥10 years) results of corneal
publicly available. Data are, however, available from the endothelial cell loss after cataract surgery. Can J Ophthalmol
authors on reasonable request. 2019;54:438‑44.
16. Bourne WM, Nelson LR, Hodge DO. Continued endothelial
Financial support and sponsorship cell loss ten years after lens implantation. Ophthalmology
1994;101:1014‑22.
Nil.
17. Reuschel A, Bogatsch H, Oertel N, Wiedemann R. Influence of
anterior chamber depth, anterior chamber volume, axial length,
Conflicts of interest and lens density on postoperative endothelial cell loss. Graefes
The authors declare that there are no conflicts of interests Arch Clin Exp Ophthalmol 2015;253:745‑52.
of this paper. 18. Yamazoe K, Yamaguchi T, Hotta K, Satake Y, Konomi K, Den S,
et al. Outcomes of cataract surgery in eyes with a low corneal
endothelial cell density. J Cataract Refract Surg 2011;37:2130‑6.
References 19. Bui AD, Sun Z, Wang Y, Huang S, Ryan M, Yu Y, et al. Factors
impacting cumulative dissipated energy levels and postoperative
1. Chen HC, Huang CW, Yeh LK, Hsiao FC, Hsueh YJ, Meir YJ, et al. visual acuity outcome in cataract surgery. BMC Ophthalmol
Accelerated corneal endothelial cell loss after phacoemulsification 2021;21:439.

Taiwan J Ophthalmol - Volume 14, Issue 1, January-March 2024 87