Fifteen-year follow-up of implantable collamer lens surgery in high myopia: Complications, cataract surgery outcomes, and ultrastructural analysis

Abstract

Purposes

This study reports 15-year outcomes of implantable collamer lens (ICL) surgery in patients with high myopia, analyzing complications, efficacy of cataract surgical, and ultrastructural mechanisms of anterior subcapsular cataract (ASC).

Methods

This retrospective study included 1208 eyes (ICL V4) and 6258 eyes (ICL V4c) from Eye & ENT Hospital of Fudan University, Shanghai, China. Follow-ups occurred preoperatively, at 1/3/6 months, annually post-ICL, and 1 month post-cataract surgery. Lens opacity, ASC rates, and complications were evaluated. Anterior capsule ultrastructure was assessed using scanning/transmission electron microscopy (SEM/TEM).

Results

The ICL V4 group (follow-up: 23.12 ± 32.06 months) exhibited higher complication rates than V4c (8.95 ± 17.40 months): lens opacity (6.46% vs. 0.16%), ASC (2.32% vs. 0.10%), and bilensectomy (3.89% vs. 0.16%). Bilensectomy improved corrected distance visual acuity from 0.56 ± 0.39 to 0.34 ± 0.51 logMAR (P = 0.006), restoring pre-ICL levels (P = 0.310), with safety/efficacy indices of 1.08 ± 0.11 and 0.99 ± 0.08. Myopia-related complications included myopic macular atrophy (0.38%), choroidal neovascularization (0.24%), and retinal detachment (0.04%). SEM/TEM revealed lens epithelial cell proliferation, fibronectin or collagen deposition, and mitochondrial abnormalities (thickened cristae, perinuclear vacuolization) in ASC cases.

Conclusions

Super high myopia, shallow anterior chamber depth, and older age are key risk factors for post-ICL cataract. While bilensectomy restores vision, high myopia-related complications persist. ASC pathogenesis involves mechanical damage and metabolic dysfunction, evidenced by ultrastructural changes. These findings underscore the need for long-term monitoring in ICL patients with high myopia.

Highlights

• •This study provides, to our knowledge, the first ultrastructural analysis using scanning and transmission electron microscopy of cataracts developed after both ICL V4 and ICL V4c implantation.
• •The findings reveal that the pathogenesis of post-ICL cataract involves a dual mechanism of mechanical damage and metabolic dysfunction.
• •We further identify key risk factors for this complication, including super high myopia, shallow anterior chamber depth, and older age.

1. Introduction

High myopia is often accompanied by variable complications threatening visual acuity, such as myopic macular degeneration, retinal detachment, cataract, and glaucoma.^1,2 The lifelong follow-up process of patients with high myopia is necessary to safeguard their eye health. Phakic intraocular lenses (PIOLs) play a pivotal role in correcting high ametropia and are an optimal option for patients who are not suitable for corneal refractive surgeries.3, 4, 5 The advantages of PIOLs include a wide range of refractive corrections, accommodation maintenance, and reversibility. However, one concern regarding PIOLs is cataract development, especially for posterior chamber phakic intraocular lenses (PC-PIOLs).⁶ Many PC-PIOLs, such as phakic refractive lenses (Zeiss Meditec, Jena, Germany), were removed from the market because of the high incidence of cataracts.⁷ Implantable collamer lenses (ICL) are the most widely used PC-PIOLs worldwide. Many studies have confirmed the excellent biocompatibility of the material,⁸ good long-term safety, and stability of refraction9, 10, 11. However, cataract development has received considerable attention. A thorough follow-up at every stage of the patients’ lives is necessary to accurately monitor the outcomes of ICL surgery, identify the occurrence of cataracts, and track the development of myopia-related complications. This not only provides robust evidence demonstrating the safety and efficacy of ICL surgery as a workable myopia correction option, but also makes a substantial contribution to ensuring the eye health of patients with high myopia.

Cataracts after PIOLs can be classified into PIOL-unrelated cataracts, characterized by nuclear cataracts (NC) or posterior capsular cataract, and PIOL-related cataracts. Surgeons and patients pay more attention to the incidence of lens opacity after PIOLs implantation and the results of phacoemulsification. Data from several recent long-term follow-up studies suggested that the incidence of cataracts increased over time, reaching 11.4%–14.5% at 10 years.10, 11, 12, 13, 14 Over the past decade, ICL products have been updated to address these issues. The introduction of the central hole design of ICL V4c, improved the aqueous circulation and reduced the incidence of cataract almost to zero.15, 16, 17, 18 However, as highlighted in the recent Asia-Pacific consensus on PC-PIOLs, the term “cataract” in the ICL context requires careful definition, encompassing both mild anterior subcapsular opacities (ASO) often insignificant to vision and visually significant cataracts requiring extraction. This consensus confirms that holed ICL designs demonstrate a much lower incidence of both lens opacities and significant cataracts compared to non-holed designs.¹⁹ Despite this improvement, Kamiya²⁰ and Chen²¹ reported cataract incidences of 1.7% and 3.58% during a 5-year follow-up, which suggested that the central hole design did not fully prevent cataract. Risk factors of cataract after ICL implantation were reported, such as shallow anterior chamber depth (ACD), age at surgery, preoperative spherical equivalence, and surgical skill22, 23, 24, 25. Among PIOL-related cataracts, anterior subcapsular cataract (ASC) is a unique type specifically associated with the implant. Critically, few previous studies distinguish between cataract types or severity (e.g., ASO vs. vision-impairing cataract). It is therefore essential to differentiate the factors contributing specifically to the development of ASC/ASO from those leading to other types of cataracts to better inform patients preoperatively.

Regardless of the type of cataract that occurs after ICL implantation, bilensectomy (PIOL extraction combined with cataract surgery),^26,27 is a safe and efficient method to improve vision. With increasing age, the need for cataract extraction is also increasing. To date, only four studies have reported the refractive outcomes of bilensectomy28, 29, 30, 31. Of the four studies, the largest scale retrospective study reported that 87 eyes with three types of PIOLs (72 eyes with ICL) underwent bilensectomy and achieved good safety and efficacy,²⁹ some of which did not improve vision due to myopia-associated comorbidities, but their refractive outcomes before and after PIOL implantation were not reported. Other studies had similar results, but with smaller sample sizes and older ICL types.^28,30,31 Further longer-term and larger-scale studies are needed to understand whether bilensectomy improves patients’ final visual outcomes.

This study aimed to report the clinical outcomes of patients with high myopia 15 years after ICL implantation, including the rates of cataract, ASC, and other complications, as well as the visual outcomes of bilensectomy. Additionally, we assessed the risk factors of cataracts after ICL implantation, especially ASC. Laboratory scanning electron microscopy (SEM) and transmission electron microscopy (TEM) were used to examine the anterior capsule membranes to elucidate the mechanism of cataracts after ICL surgery. A comprehensive study of a large cohort spanning from pre-ICL surgery to post-bilensectomy surgery, can provide a more thorough understanding of lifelong ocular health implications for high myopia.

2. Methods

2.1. Study population

This retrospective cohort study included all patients with myopia who underwent ICL implantation at the Fudan Eye, Ear, Nose, and Throat Hospital (Shanghai, China) between Jan.1st 2006 and Sep.30th 2021. All ICL surgeries were performed by an ophthalmologist (XW) who had experience with ICLs before beginning the study. The flowchart of the study design is demonstrated in Fig. 1. The inclusion criteria were patients with a minimum follow-up duration of 1 month.

Fig. 1.

Flowchart of the 15-year outcomes following ICL implantation. This retrospective cohort study followed the reporting criteria for Strengthening the Reporting of Observational Studies in Epidemiology (STORE). This flowchart showed the study design and included patients.

Patients were excluded for contraindications of ICL implantation: central ACD <2.6 mm; endothelial cell density (ECD) count <2000 cells/mm²; refractive media opacity that significantly impairs vision; history of autoimmune conditions, including diabetes, rheumatoid arthritis, multiple sclerosis, and systemic lupus erythematosus; history of diseases affecting vision other than myopia and astigmatism, such as uveitis, glaucoma, or retinal detachment.

Indications for bilensectomy included a decreased corrected distance visual acuity (CDVA) to decimal 0.5 or less and symptoms of blurred vision caused by cataracts, even in the absence of a discernible loss in visual function.

Patients’ medical files were reviewed for sex, ICL implantation data, ICL explantation data, ICL model and size, complications during long-term follow-up, anterior segment optical coherence tomography (CASIA 2; Tomey Corp, Nagoya, Japan), Pentacam HR (Type 70900; Oculus Optikgeräte GmbH, Wetzlar, Germany), and slit-lamp biomicroscopy to confirm the cataract type (ASC, posterior subcapsular cataract (PSC), or NC). The classification of lens opacity was purely clinician-based.

2.2. Preoperative and postoperative assessment

The visual outcomes (CDVA, uncorrected distance visual acuity (UDVA), and manifest refraction) were available at the following timepoints: before and 1 month after ICL implantation, 1 month before and after bilensectomy.

The other anterior segment measurements were conducted before ICL implantation, including axial length (intraocular lens (IOL) Master, Carl Zeiss, Germany), corneal topography (Pentacam HR, Type 70900; Oculus Optikgeräte GmbH, Wetzlar, Germany), intraocular pressure (IOP, Tonemeterx-10, Canon, Japan), sulcus to sulcus (UBM, Quantel Medical, France), and vaults (Pentacam HR, Type 70900; Oculus Optikgeräte GmbH, Wetzlar, Germany).

2.3. Phakic IOL size and power calculation

The appropriate power and size of the ICL were selected for implantation surgery. ICL power calculation was performed using STAAR Surgical formula based on subjective refraction, keratometry, and ACD. In all eyes, emmetropia was targeted to correct refractive error. The ICL size was selected based on the manufacturer's instructions. An experienced technician prescribed the recommended ICL size according to the UBM measurements and STAAR Surgical recommendations based on the white-to-white horizontal corneal diameter and ACD. The lens models used were the ICL V4, Toric ICL V4, ICL V4c, and Toric ICL V4c.

2.4. Intraocular lens power calculation

For IOL power calculation, it was calculated preoperatively using the SRK-T formula based on ocular biometry (IOL Master, Carl Zeiss Meditec AG). The targeted spherical equivalent (SE) was then recorded.

2.5. Surgical procedure

All the surgeries were performed by one ophthalmologist (XW). The ICL implantation and surgical procedures were the same as those used in our previous studies. At the temporal corneoscleral limbus, a 3-mm incision was made. After injection of a viscoelastic surgical agent (1.7% sodium hyaluronate, Bausch & Lomb, China) into the anterior chamber to preserve the ACD, an EVO ICL was placed into the anterior chamber using an injector cartridge. Occasionally, the ICL was then topped with an ophthalmic viscosurgical device (OVD), and the four haptics of the ICL were swept beneath the iris using an ICL positioning tool. The OVD was then irrigated with a balanced salt solution. Lastly, the incision was closed by the surgeon using a cotton swab to gently push the cornea.

The bilensectomy (ICL extraction and cataract surgery) was as follows. After the pupil was fully dilated, two side ports were created so that an OVD was injected into the anterior chamber. A 3.1-mm temporal main incision was created. All four footplates were then manipulated anterior to the iris, and the ICL was removed through the main incision. Once the ICL was removed, the OVD was added to the anterior chamber, and a continuous curvilinear capsulorhexis was created using forceps. Subsequently, standard phacoemulsification was performed occasionally, and then the capsular tension ring and IOL were implanted into the capsular bag. Intracameral cefuroxime was administered at the end of the surgery, and the incisions were hydrated for closure. All surgeries were performed under topical anesthesia.

2.6. Ultrastructural analysis of the anterior capsule

Four anterior lens capsules and four ICLs were examined by SEM. Six anterior lens capsules were examined by TEM.

Immediately after removal, anterior lens capsule specimens were fixed overnight (2.5% glutaraldehyde and 2.0% paraformaldehyde), osmicated, and dehydrated through an ascending graded ethanol series.

For SEM, specimens were dried in a Hitachi HCP-2 (Tokyo, Japan) critical-point drying apparatus and coated with platinum in an ion-sputtering machine (LEICA EM AG600, Germany). ICLs were cut in half to observe the front and back sides. Afterward, the specimens were examined under a SEM (Hitachi SU8010, Japan).

For TEM, specimens were embedded in Epon overnight and polymerized at 60°C for 48 h. Ultrathin sections of 30–50 nm were obtained and stained with uranyl acetate and lead citrate, and then examined by TEM (Philips CM120; Holland).

2.7. Statistics

The patient characteristics and visual outcomes at pre-ICL, post-ICL, pre-bilensectomy, and post-bilensectomy are summarized as means and standard deviations for continuous variables or counts and percentages for categorical variables. For survival analysis, the time from ICL implantation until the main outcomes were first noted was calculated; if no main outcome was observed for eyes with less than the maximum follow-up duration available, the data were censored at the last available follow-up. Mixed-effects Cox regression was used to calculate the covariate-adjusted hazard ratios and 95% confidence intervals (CIs) of ASC-free survival. The CDVA logMAR at follow-up was compared using a generalized linear mixed-effects model. Patient identifiers and study eyes were entered into all the models (including the mixed-effects Cox model and generalized linear mixed-effects model) to account for inter-eye correlations. All statistical analyses were performed using SPSS software (version 27, SPSS Inc., IBM, USA). Data analysis was performed between July 1, 2023, and October 31, 2023.

3. Results

Finally, we included 1208 eyes of 651 patients who underwent ICL V4 implantation and 6258 eyes of 3175 patients who underwent ICL V4c implantation. Table 1 shows the baseline characteristics and clinical outcomes before and after ICL implantation in the ICL V4 and V4c groups. The ages of ICL implantation were 29.04 ± 7.51 (17–51) years and 27.77 ± 6.64 (17–54) years in the ICL V4 and V4c groups. After ICL surgery, UDVA was 0.20 ± 0.23 logMAR in the ICL V4 group and −0.01 ± 0.13 logMAR in the V4c group. The postoperative CDVA was 0.08 ± 0.17 logMAR and −0.03 ± 0.10 logMAR in the ICL V4 and V4c groups, respectively. The maximum follow-up time was 182.53 months for the ICL V4 group and 96.00 months for the ICL V4c group.

Table 1.

Preoperative patient demographics.

	ICL V4 (1208 eyes of 651 patients)				ICL V4c (6258 eyes of 3175 patients)
	Mean/n	SD/%	Min	Max	Mean/n	SD/%	Min	Max
Sex (No. %)
female	799	66.14%			4897	78.25%
male	409	33.86%			1361	21.75%
ICL model (No. %)
Spherical ICL	887	73.43%			2641	42.20%
Toric ICL	321	26.57%			3617	57.80%
Before ICL implantation
Age (Years)	29.04	7.51	17.00	51.00	27.77	6.64	17.00	54.00
CDVA (LogMAR)	0.22	0.25	−0.08	1.30	0.02	0.11	−0.18	1.30
Axial length (mm)	29.49	2.11	20.61	36.72	27.40	1.68	22.38	36.78
Spherical Equivalence (D)	−15.50	4.76	−31.00	14.88	−9.80	3.61	−29.37	−1.50
ACD (mm)	3.19	0.24	2.29	4.11	3.21	0.24	2.77	4.50
WTW (mm)	11.80	0.33	10.70	13.20	11.62	0.37	10.10	13.00
After ICL implantation
Follow-up time (months)	23.12	32.06	1.00	182.53	8.95	17.40	1.00	96.00
post ICL UDVA (LogMAR)	0.20	0.23	−0.18	2.00	−0.01	0.13	−0.30	1.40
post ICL CDVA (LogMAR)	0.08	0.17	−0.18	2.00	−0.03	0.10	−0.18	1.00
Safety index	1.03	0.04	0.63	1.35	1.01	0.02	0.96	1.27
Efficacy index	1.01	0.04	0.63	1.33	1.01	0.02	0.84	1.27
Vault (um)	456.92	176.40	0.00	1170.00	516.67	205.33	0.00	1560.00

n, number; SD, standard deviations; ICL, implantable collamer lens; PC-ICLs, posterior chamber phakic intraocular lenses; CDVA, corrected distance visual acuity; UDVA, uncorrected distance visual acuity; ACD, anterior chamber depth; WTW, white to white.

3.1. Overall complications after ICL implantation

Table 2 demonstrated the complications after ICL implantation. Twenty-two eyes (1.82%) in the ICL V4 group and six eyes (0.10%) in the ICL V4c group developed myopic macular atrophy; seven eyes (0.58%) in the ICL V4 group and 11 eyes (0.18%) in the ICL V4c group developed myopic choroidal neovascularization. In addition, a total of seven eyes underwent ICL realignment due to trauma causing lens dislocation rotation. Of these seven eyes, one eye developed a cataract 142.10 months postoperatively, which was excluded from the statistical analysis thereafter.

Table 2.

Complications after ICL implantation.

	ICL V4 (1208 eyes of 651 patients)			ICL V4c (6258 eyes of 3175 patients)
	nA	Rate	Related outcomes	nA	Rate	Related outcomes
Cataract
ASC	28	2.32%	47 eyes underwent bilensectomy	3	0.05%	10 eyes underwent bilensectomy
PSC/NC	50	4.14%		7	0.11%
Myopia-related
Myopic macular atrophy	22	1.82%	CDVA (logMAR) = 0.076 ± 0.62 (0–0.22), no eye loss of CDVA; one eye with a lamellar macular hole, CDVA Senllen = 20/20	6	0.10%	CDVA (LogMAR) = 0.022 ± 0.59 (−0.08–0.22), two eyes with macular schisis loss of CDVA
Myopic choroidal neovascularization or hemorrhage	7	0.58%	CDVA decimal = 0.46 ± 0.25 (0.15–0.6), three eyes loss of CDVA	11	0.18%	CDVA decimal = 0.85 ± 0.34 (0.2–1.2), five eyes loss of CDVA
Retinal detachment	2	0.17%	before ICL, CDVA (Snellen) = 20/25,20/50; post ICL, CDVA (Snellen) = 20/16, 20/20; post RD, CDVA (Snellen) = 20/200, 20/400	1	0.02%	before ICL, CDVA Snellen = 20/32; after ICL, CDVA (Snellen) = 20/25; after RD, CDVA (Snellen) = 20/40
Surgery-related
Pigment dispersion	1	0.08%	after ICL CDVA (Snellen) = 20/25; Vault 500 μm; developed ASC later	0	0.00%
Ocular Hypertension or Glaucoma	2	0.17%	after ICL, CDVA (Snellen) = 20/20, 20/25, IOP = 39, 40 mmHg; after treatment, CDVA (Snellen) = 20/20, 20/32, IOP = 21.6, 21.6 mmHg	3	0.05%	after ICL, CDVA (Snellen) = 20/12.5, 20/12.5, 20/20; IOP = 19.7, 18.0 mmHg (LASIK history), 11.4 mmHg (after treatment)
Traumatic dislocation	7	0.58%	after alignment CDVA (Snellen) = 20/63, 20/20, 20/25, 20/25, 20/25, 20/32,vault = 300–450 μm; one with ASC, after alignment CDVA (Snellen) = 20/32, vault = 130 μm;	0	0.00%	-
ICL extraction or realignment	5	0.41%	after extraction or realignment, CDVA (Snellen) = 20/20, 20/25, 20/20, 20/20, 20/20,Vault = 290, 1000, 900, 650, 760 μm	0	0.00%

A, the number described eyes; ICL, implantable collamer lens; CDVA, corrected distance visual acuity; IOP, intraocular pressure measured by non-contact tonometer; ASC, anterior subcapsular cataract; PSC, posterior subcapsular cataract; NC, nuclear cataract.

One case of pigment dispersion occurred two years after ICL implantation. The patient was diagnosed with hyperthyroidism 10 years after ICL implantation; this case was also excluded from the following analysis.

One eye in the ICL V4 group had an elevated intraocular pressure, which was controlled with IOP-lowering treatment. Three eyes of two patients with ICL V4c developed open-angle glaucoma. All three eyes were treated with drugs until the final follow-up.

Five eyes underwent ICL exchange or realignment because of inappropriate lens size or lens rotation. All five achieved ideal visual outcomes and vaults, and no complications occurred.

3.2. Subgroup analysis of eyes developing ASC

Of the 7466 eyes included in this study, 78 eyes (6.46%) in the ICL V4 group and 10 eyes (0.16%) in the ICL V4c group developed lens opacity during the long-term follow-up. Of the 88 eyes developing lens opacity, 47 eyes (3.89%) in the ICL V4 group and 10 eyes (0.16%) in the ICL V4c group underwent bilensectomy at a mean of 121.90 ± 33.08 months postoperatively. In the cases of lens opacity, the incidence of ASC was 2.32% (28 of 1208 eyes) and 0.05% (3 of 6258 eyes) in the ICL V4 and V4c groups, respectively (Table 2).

Fig. 2 shows the Kaplan-Meier curves for the development of lens opacity, ASC, and cataract extraction over time, with the corresponding at-risk table provided below. The estimated median survival times were 135.77 ± 14.38 (95% CI: 107.58–163.96) months, 157.54 ± 4.46 (95% CI: 148.79–166.28) months, and 142.07 ± 4.54 (95% CI: 133.37–150.76) months.

Fig. 2.

Kaplan-Meier curves for the rate of lens opacity, ASC, and cataract extraction over time. The black line shows the estimated percent at risk, and the blue line shows the 95% CI.

3.3. Clinical outcomes and complications of bilensectomy

Further analysis revealed that 57 eyes underwent bilensectomy because of clinically significant cataracts. The results showed 53 eyes in the follow-up records. Four eyes of three patients were lost to follow-up due to COVID-19. The baseline values of these four eyes were not significantly different from those of the other eyes. Table 3 shows the patient characteristics and visual outcomes before and after ICL implantation and bilensectomy. The safety and efficacy index of bilensectomy was 1.08 ± 0.11 and 0.99 ± 0.08, respectively. The CDVA was improved from 0.56 ± 0.39 logMAR to 0.34 ± 0.51 logMAR after bilensectomy (P = 0.006, Table 3), and the post-bilensectomy CDVA was statistically equal to that of the pre-ICL CDVA (P = 0.31, Table 3). Fig. 3A shows the CDVA change lines, and Fig. 3B shows the cumulative CDVA at all follow-up times. Forty-seven (83%) eyes had Snellen visual acuity of 20/40 or higher after bilensectomy (Fig. 3A). Compared with the pre-bilensectomy CDVA, 67.92% of eyes improved by at least one line, and 90.57% did not lose lines (Fig. 3B). Five eyes lost one or more lines, one eye lost one line, and four eyes lost two or more lines (Fig. 3B).

Table 3.

Clinical outcomes and complications of bilensectomy (57 eyes underwent bilensectomy).

	Total				ASC				PSC/NC/CC
	Mean /n	SD /%	Min	Max	Mean /n	SD /%	Min	Max	Mean /n	SD /%	Min	Max
Sex (No. %)
female	35	59.32%			16	84.2%			19	50.00%
male	22	37.29%			3	15.8%			19	50.00%
ICL model (No. %)
Spherical ICL	46	79.31%			14	73.68%			32	84.21%
Toric ICL	11	18.97%			5	26.32%			6	15.79%
Before and after ICL implantation
Age of ICL implantation (Years)	35.78	7.49	19.00	50.00	33.74	8.19	19.00	50.00	37.30	7.22	20.00	48.00
Axial length (mm)	31.46	2.26	26.46	35.45	30.86	2.22	28.06	35.45	31.95	2.23	26.46	36.02
Spherical Equivalence (D)	−19.81	4.86	−28.00	−12.63	−19.11	4.65	−28.13	−12.63	−19.83	6.50	−28.00	0.00
ACD (mm)	3.12	0.22	2.74	3.56	3.12	0.27	2.80	3.56	3.11	0.20	2.74	3.50
WTW (mm)	11.73	0.31	11.00	12.30	11.80	0.32	11.20	12.45	11.74	0.32	11.00	12.30
pre ICL CDVA (LogMAR)	0.40	0.30	0.00	1.30	0.25	0.21	0.00	0.82	0.48	0.32	0.00	1.30
post ICL UDVA (LogMAR)	0.55	0.32	0.10	1.50	0.25	0.25	−0.08	0.82	0.38	0.26	−0.08	1.00
post ICL CDVA (LogMAR)	0.20	0.23	−0.18	0.82	0.05	0.14	−0.08	0.52	0.26	0.24	−0.18	0.82
Safety index	1.05	0.04	0.97	1.14	1.04	0.02	1.00	1.07	1.05	0.04	0.97	1.16
Efficacy index	1.01	0.04	0.94	1.13	1.00	0.03	0.94	1.05	1.02	0.05	0.94	1.16
Vault	298.72	170.11	90.00	780.00	292.22	136.28	90.00	540.00	402.00	194.61	90.00	800.00
Before and after bilensectomy
Age of bilensectomy (Years)	45.58	7.41	27.00	59.00	43.37	8.08	27.00	59.00	46.89	7.26	30.00	59.00
ICL duration (months)	121.90	33.08	18.00	171.00	118.70	31.58	57.87	159.60	119.44	35.80	18.93	171.73
pre-IOL UDVA (LogMAR)	0.81	0.32	0.30	1.00	0.44	0.33	0.00	1.00	0.62	0.27	0.10	1.00
Pre-IOL CDVA (LogMAR)	0.56	0.39	2.00	0.10	0.62	0.47	0.10	2.00	0.58	0.35	0.00	2.00
Post-IOL UDVA (LogMAR)	0.31	0.23	0.04	1.00	0.44	0.33	0.00	1.00	0.62	0.27	0.10	1.00
Post-IOL CDVA (LogMAR)	0.34	0.51	−0.18	3.00	0.22	0.49	−0.18	2.00	0.42	0.52	0.00	3.00
Safety index	1.05	0.11	0.50	1.35	1.10	0.11	0.95	1.35	1.03	0.11	0.50	1.15
Efficacy index	0.99	0.08	0.85	1.17	1.01	0.07	0.86	1.17	0.98	0.08	0.85	1.14
Complications after bilensectomy
PCO	5	9.43%	After YAG laser treatment, CDVA = 0.5, 0.6, 0.8, 0.4, 0.25, 0.8, 1.0, one eye lost CDVA due to CNV
Retinal detachment	1	1.89%	After ICL, CDVA = 0.6; before ICL, CDVA = finger count; after ICL, CDVA = 0.9; after vitrectomy, CDVA = 0.1
Macular retinoschisis	1	1.89%	After ICL, CDVA = 0.8; before ICL, CDVA = 0.25; after ICL, CDVA = 0.4; after vitrectomy, CDVA = 0.2
Endothelium cell loss	1	1.89%	Before ICL, ECD = 3072, before ICL, CDVA = 0.5, ECD = 1905, vault = 200; 1 week after ICL, CDVA = 0.7; 2 years after ICL, CDVA = 0.5, ECD = 1095

ICL, implantable collamer lens; ASC, anterior subcapsular cataract; PSC, posterior subcapsular cataract; NC, nuclear cataract; SD, standard deviations; IOL, intraocular lens; CDVA, corrected distance visual acuity; UDVA, uncorrected distance visual acuity; ACD, anterior chamber depth; WTW, white to white; ECD, endothelium cell density; PCO, posterior capsular opacity.

Fig. 3.

Visual outcomes at all follow-up times after ICL surgery. A. Cumulative Snellen visual acuity at all follow-up times. B. CDVA logMAR change lines at all follow-up times. CDVA = corrected distance visual acuity, ASC = anterior subcapsular cataract, PIOL = phakic intraocular lens. The corresponding numerical values for the bar graph are presented in Supplementary Table 1, 2.

The rate of posterior capsular opacity (PCO) was 9.8% (5 eyes of 3 patients), occurring at a mean of 19.48 ± 8.89 (8.77–33.47) months after bilensectomy. All these cases were successfully treated with neodymium: YAG (Nd: YAG) laser capsulotomy (Table 3).

The other myopic-related complications were as follows. One eye had foveoschisis 6 months after a bilensectomy and lost one line of CDVA after posterior vitrectomy. Another eye experienced postoperative retinal detachment and lost 10 lines of CDVA. Three eyes of two patients had corneal edema, with follow-up data only one day post bilensectomy due to the COVID-19 epidemic (Table 3). One case of endothelial cell loss 1 year after bilensectomy was observed. The ECD declined from preoperative 1905 cells/mm² to postoperative 1095 cells/mm², and the BCVA was Snellen 20/40, the same as before bilensectomy (Table 3).

3.4. Risk factors of lens opacity, ASC, and bilensectomy

Among all patients implanted with ICL V4c, three eyes of two patients developed ASC and underwent bilensectomy. The patients’ preoperative characteristics are shown in Supplementary Table 3 (Patient No.2 and No.6). Both patients had a preoperative refractive error exceeding −16.00 D, were over 35 years old at the time of ICL implantation, had ACD ranging from 2.80 to 2.98 mm, and a relatively low pre-bilensectomy vault of 100–200 μm.

Supplementary Table 4 summarizes the hazard ratios for lens opacity, ASC development, and bilensectomy using Cox regression. There was no evidence that vault influenced lens opacity (P = 0.27, HR = 0.999, 95% CI: 0.997–1.001), ASC development (P = 0.662, HR = 0.995, 95% CI: 0.995–1.000), and cataract surgery (P = 0.809, HR = 0.997, 95% CI: 0.997–1.003), but eyes with lower ACD were more likely to develop lens opacity (P = 0.006, HR = 0.138, 95% CI: 0.034–0.56) and undergo bilensectomy (P = 0.001, HR = 0.036, 95% CI: 0.005–0.262). Older age at ICL implantation was also associated with increased risk of lens opacity (P = 0.001, HR = 1.059, 95% CI: 1.023–1.097) and undergoing bilensectomy (P = 0.015, HR = 1.065, 95% CI: 1.012–1.121) during long-term follow-up. Lower pre-ICL SE (higher myopia) was the only factor related to a higher risk of ASC (P = 0.04, HR = 0.88, 95% CI: 0.779–0.994).

3.5. Ultrastructure findings of different types of cataract

Fig. 4 demonstrated the SEM morphology of the anterior lens capsule. In the specimen from a patient with NC after ICL surgery, most of the lens epithelial cells (LECs) immediately adjacent to the basement membrane have typical epithelial cell characteristics, approximately 10 μm wide. At high magnification, a few filaments approximately 8-10 μm long, can be seen between and on the surface of the LECs (Fig. 4A). Fig. 4B–D demonstrated the SEM results of specimens from three patients with ASC after ICL implantation. Fig. 4B showed the monolayered lens epithelium is covered with a layer of proliferating fibers spanning several cells up to 18–20 μm. Fibers are also present on the surface of the complex epithelium, and it is not well defined due to the edematous cells (Fig. 4C). On the cell surface at the junction of the compound LECs and the monolayer LECs, two flat, elongated, spindle-shaped cells (about 40 μm long and 20 μm wide) can be found, but without protrusions that are closely related to the collagen fibrils (Fig. 4D).

Fig. 4.

Morphology of lens epithelium cells in different types of cataract observed by SEM. All the clinical data of the samples were demonstrated in Supplementary Table 3. A. (Left eye of Patient No. 2) Multiple layers of LECs from a patient with nuclear cataract with a small amount of fibrous materials at higher magnification. B. (Left eye of patient No. 3) A single layer of LECs from a patient with ASC. Longer fibrous material, crossing several cells were observed at higher magnification. C. (Left eye of patient No. 1) Multiple layers of LECs from a patient with NC with fibrous material. LECs were edema due to the surgical damage. D. (Right eye of patient No. 4) Two flattened, elongated,spindle-like cells at the junction of multilayer LECs and monolayer LECs. ➩ Solid arrow = fibrous materials, ○ red circle = spindle-like cells. SEM = scanning electron microscopy, NC = nuclear cataract, ASC = anterior subcapsular cataract, LECs = lens epithelial cells.

Fig. 5 demonstrated TEM results of the anterior capsules from patients with NC (Fig. 5A), NC + ASC (Fig. 5B), and ASC (Fig. 5C, D). The specimen from the patient with NC showed monolayer cuboidal epithelial cells (10–15 μm) with huge horizontal spacing between adjoining epithelial cells. Due to degradation of cellular function, cell atrophy, thinning, and vacuolization were observed (Fig. 5A). In another patient with NC and ASC (predominant NC), TEM revealed abnormal epithelial cells with morphological changes from cuboidal to triangular and flattened shapes, as well as obvious intercellular vacuolization (Fig. 5B). Long oval monolayer epithelial cells were observed in the LECs of a patient with ASC (Fig. 5C). The LECs showed light cytoplasmic staining and sparse cytoplasm, with vacuolization located mainly in the intracellular portion close to the nucleus (Fig. 5C). High-magnification microscopy showed the characteristic changes in the morphology of the mitochondria, which were more numerous, with enlarged and densely packed cristae (Fig. 5C). A cell surface covered with collagen-like filaments was observed in another patient with ASC (Fig. 5D), which corresponded to the filaments in the SEM image (Fig. 4). In addition, intracellular and intercellular vacuolization, and loose epithelial cell-to-basal surface connections (Fig. 5D).

Fig. 5.

Morphology of lens epithelium cells in different types of cataract observed by TEM. All the clinical data of the samples were demonstrated in Supplementary Table 3. A. (Right eye of patient No. 8) Monolayer cubic LEC in NC with horizontal spacing between adjoining LECs. Vacuolization is located primarily at cell junctions. B. (Left eye of patient No. 7) Multilayer flattened LECs in NC, with vacuolization between adjoining cells. C. (Left eye of patient No. 6) Flattened monolayer LECs in ASC, with vacuolization around the nucleus, and larger numbers of enlarged rough endoplasmic reticulum. Higher magnification demonstrated an increase in the number of cistens. D. (Left eye of patient No. 4) Collagen covering the surface of LEC in ASC. Large amount of intracellular and intercellular vacuolization was observed. ★ Pentagrams = large vacuolization between cells; ∗ small asterisks = small perinuclear vacuolization; → line arrows = cistens; ➩ Solid arrows = fibrous materials. TEM = transmission electron microscopy, LECs = lens epithelial cells, NC = nuclear cataract, ASC = anterior subcapsular cataract.

4. Discussion

In this large-scale cohort with a maximum follow-up of 15 years after phakic refractive surgery, the incidence of lens opacity after ICL V4 and ICL V4c implantation was 6.46% and 0.16%, respectively, while the incidence of ASC was 2.32% and 0.10% in the ICL V4 and ICL V4c groups. The proportion of eyes undergoing bilensectomy implantation due to lens opacity was 3.89% and 0.16% in the ICL V4 and ICL V4c groups, respectively. The estimated median survival times of lens opacity, ASC, and bilensectomy after ICL V4 implantation were 135.77 ± 14.38 months, 157.54 ± 4.46 months, and 142.07 ± 4.54 months.

The findings of this study complement observations made in previous investigations of ICL implantation by describing the incidence of lens opacity, ASC, and other ocular complications associated with high myopia, as well as the outcomes of cataract surgery during whole-life follow-up. The results of our study were compared with previous projections in Supplementary Table 5. In a three-year FDA trial, Sanders reported that 14 (2.7%) eyes had ASC, and two eyes (0.4%) had clinically significant cataracts.³² In the same report, the estimated cumulative probability over the 7-year follow-up was 7%, whereas 31 (5.9%) of 526 eyes had ASC.²⁵ Alfonso et al. reported an ASC incidence of 0.94% (32/3420) and a bilensectomy incidence of 0.61% (21/3420); all the eyes underwent ICL V4.³³ The follow-up periods of the three ICL models were different; thus, the rate of clinically significant cataracts in the ICL V4c group may have been underestimated. Guber et al. reported that the rates of lens opacity were 31% at 2 years, 41% at 5 years, 51% at 8 years, and 55% at 10 years.¹¹ This report attributed the high incidence to their examination method: wider view and detection of peripheral lens opacity under full dilation. The rate of phacoemulsification (18% after 10 years) was similar to Schimidinger's³⁴ (17% after 10 years), higher than ours. This rate was high, possibly owing to the long study period. Nakamura reported an ASC incidence of 11.4% in eyes with ICL V4, and 3.51% of eyes underwent cataract extraction in 10-year follow up,¹⁰ which was similar to our results. The reported incidence of ASC in ICL V4c is 0%–3.85%.^9,20,21,35 Our previous ICL V4 and V4c comparative study reported that the incidence of lens opacity was 3.84% in total, that of ASC was 3.85% in the V4 group and 1.28% in the V4c group. Moreover, 2.56% (2/78) of eyes in the V4 group underwent bilensectomy during five years of follow-up.²¹ Another study on high myopia reported an ASC incidence of 3.61% after ICL V4c implantation, but none was clinically significant at five years.³⁵ Actually, 3.89% of patients in the ICL V4 group and 0.16% of patients in the ICL V4c group underwent cataract extraction in the current study. ASC is a surgery-related cataract, as distinct from age- and myopia-related cataracts. The percentage of ASC in our total population with lens opacity was close to that reported by Vargas et al. (ASC accounted for 43% of all cataract types).²⁹ Moreover, we believe that as ICL products are upgraded and surgical techniques are improved, the incidence of post ICL cataracts will decline further.

The present study expands upon the findings from the earlier studies by differentiating the risk factors of cataract and ASC development. We found that preoperative SE, age at ICL surgery, and shallow ACD are associated with increasing risk of lens opacity. This is consistent with Gimbel.²² Sanders stated that preoperative myopia (>12.00 D) and patient age (>40 years) were significant factors in the development of ASC.²⁵ Maeng's results showed that the ASC rate in those with a low vault was obviously higher than the overall rate and suggested an optimum cutoff of 51.7 μm of vault, 45 years of age, and −14.00 D to prevent cataract development.²³ For those with shallow ACD (<2.8 mm), Lim observed that 11.1% (2/18) of eyes developed ASC at 8–11 months after ICL V4.³⁶ However, ICL V4c implantation in patients with high myopia and shallow ACD showed no evidence of cataract during 15.35 months (15–25 months) of follow-up.³⁷ This suggests aqueous dynamics improvements from central hole designs may mitigate anatomical risks. Mechanistically, aqueous humor alterations in high myopia involve amino acid, carbohydrate, and lipid metabolism,³⁸ matrix metalloproteinases³⁹ and TGF-β2 levels.⁴⁰ This should be the focus of future studies. We speculate that metabolic factors underlie high myopia's role in the early development of cataracts, warranting further study.

Vault is an intermediate variable influenced by ACD, WTW, and other factors. We observed that the vault was not associated with ASC or cataract development, which may be due to the small number of low-vault cases. Whether the vault is correlated with the incidence of cataract has been inconsistent in previous studies. Schmidinger observed that a lower vault was associated with ASC in 10 years of follow-up.¹³ However, another 10-year follow-up of patients with low and normal vault showed no cases that developed ICL-induced ASC in either groups.⁴¹ Another case report of persistent low vault also showed no evidence of crystalline lens opacity or a decrease in UDVA during the 6-year follow up.⁴² These suggested that the central hole prevents the potential risk associated with a low vault, and ASC formation is affected by high myopia, age at surgery, and ACD. These findings align with the Asia-Pacific consensus, confirming low vault as a known risk factor for lens opacities (including ASC), though it may not necessitate ICL removal. When visually significant cataracts develop despite these advances, the consensus notes that bilensectomy becomes necessary. Reported bilensectomy rates range from 0.5% to 5.5% at 5–10 years to 7.6%–14.3% beyond 10 years,¹⁹ underscoring the importance of long-term monitoring.

Since ICL V4c, with a central hole, entered the market, the incidence of ASC and cataracts has shut down. The aqueous fluid penetrates the central hole so that the crystalline lens can maintain metabolism and transparency, even when there is no vault and nearly contact.^16,17 However, three eyes of two patients at our center developed clinically significant ASC at 57.87–78.47 months after ICL V4c implantation. Kamiya et al. reported three eyes with ASC after implantation of the ICL V4c.²⁰ All the cases in our study and two cases in Kamiya's study had similar characteristics: age at surgery >35 years, with a preoperative SE less than −15 D. Although their vaults were in the range of 250–350 μm within two years postoperatively, all of them had vaults about 100 μm at the time of ASC occurrence. ASC developed in the middle peripheral part of the crystalline lens, appearing in a ring-like pattern corresponding to the ICL contact zone (as has been reported in previous literature). This suggests that a low ICL vault can disrupt aqueous fluid circulation, impairing lens metabolism. Specifically, for V4c lenses, where central aqueous flow is theoretically preserved, the peripheral stagnation or flow gradient might still contribute to ring-pattern ASC due to localized metabolic compromise.

LECs form a critical barrier between the aqueous humor and lens fibroblasts in ASC pathogenesis. ASC is associated with a variety of causes, such as trauma, surgery,⁴³ herpes keratitis uveitis,⁴⁴ and retinitis pigmentosa.⁴⁵ Long-term ICL effects on human LEC remain poorly understood due to the limited sample sizes and follow-up (5 years).^25,46

Our TEM/SEM analysis revealed key ASC features in post-ICL patients: abnormal proliferation with scattered spindle-shaped cells, epithelial-mesenchymal transition, and extracellular fibrotic deposits containing collagen and fibronectin. These findings align with established ASC mechanisms involving LEC metaplasia and basement membrane remodeling. 47, 48, 49, 50, 51

Notably, we identified distinctive vacuolization patterns: severe ASC cases showed predominant perinuclear vacuoles, contrasting with large intercellular vacuoles in nuclear cataracts. Additionally, this progression correlated with thickened and rough mitochondria cristae. Similar findings were observed in porcine models.^15,52

The observed perinuclear vacuolization represents a novel human ASC marker unreported in prior models. Given that oxidative stress has an established role in cataractogenesis⁵³ and antioxidants like GSH, vitamin E/C, and quinone has been proven to slow cataract formation,⁵⁴ our findings suggest antioxidant strategies may mitigate post-ICL ASC development.

In one ASC patient, we observed a monolayer of LECs—similar to post-traumatic ASC where basement membrane regeneration replaces damaged LECs.⁵⁵ Notably, a patient with a low vault of 90 μm at ASC onset required cataract surgery six years later, despite maintaining 20/40 corrected vision. This slow progression suggests mechanical friction from the low vault as the primary trigger; the potential contribution from metabolic dysfunction may lead to ASC development. This hypothesis needs to be proven in future studies.

Few preceding studies focused on the outcomes of ICL extraction and cataract surgeries,28, 29, 30 we found that retinal comorbidities associated with high myopia may limit cataract surgery outcomes. The CDVA improved from 0.56 ± 0.39 to 0.34 ± 0.51, which reflected slightly inferior numerical results to values previously reported28, 29, 30 (0.27 logMAR or better). This is likely due to the higher degree of myopia (pre-ICL SE: −19.81 ± 4.86 D) compared with −7.99 D to −16.4 D in other long-term follow-up studies, resulting in eyes with inferior visual potential. In other words, the safety index was 1.08 ± 0.11, a total of 83% of the eyes had Snellen visual acuity of 20/40 or better, and half of the eyes achieved better BCVA than pre-ICL. This was comparable to previous studies28, 29, 30, which reflected satisfactory efficacy and safety. Although a large proportion of eyes can improve visual acuity after cataract surgery, approximately 1/10 of them did not because of the concomitant retinal comorbidities, such as myopic macular atrophy (20 eyes, 35.09%), CNV (4 eyes, 7.02%), and RD (1 eye, 1.75%), all complications associated with high myopia. This result suggests that it is desirable for patients with high myopia to retain their vision through cataract surgery. However, high myopia-related retinal comorbidities may restrict the results of cataract surgery. Ophthalmologists should fully communicate with patients preoperatively.

One case of RD and one case of macular foveoschisis occurred after bilensectomy. Vargas et al. also reported one case of postoperative RD after cataract surgery with ICL removal, but they attributed this to high myopia and younger age at surgery (44.4 years).²⁹ Intra ocular surgery has been associated with an increased risk of RD, with an incidence of RD of approximately 1.0% after cataract surgery, compared with 2.2% in patients with high myopia.⁵⁶ A younger age of 40–54 years (HR = 5.22, compared with age >75 years) and high myopia (HR = 6.12) were also risk factors associated with an increased risk of RD.⁵⁷ Several recent studies confirmed that phakic intraocular implantation surgery does not increase the incidence of RD.58, 59, 60 Thus, we believe that the higher degree of myopia and relatively younger age (mean age at cataract development: 45.58 ± 7.41 years old) in our study population increased the risk of RD.

In our study, ECC loss occurred in one eye, which has not been reported in previous studies28, 29, 30. There is no evidence that this surgical maneuver significantly damages the endothelium; thus, we believe that it is a result of a surgical accident.

PCO occurred in 9.8% of eyes (5 eyes) in our study at approximately 19.48 ± 8.89 (8.77–33.47) months postoperatively, and all of these PCOs were successfully managed with laser, with no patients experiencing intraoperative posterior capsule rupture. Meier reported that PCO occurred in 15 eyes (39%), at a mean of 2.9 ± 2.2 years postoperatively.³⁰ Vargus reported a PCO incidence of 4/87 (4.6%) at a mean of 14.5 ± 5.6 months of follow-up.²⁹ There is no evidence whether ICL predisposes individuals to PCO, but high myopia predisposes individuals to PCO.

This study has some limitations. Firstly, we have not been able to determine whether cataracts have actually occurred in patients who have not yet been followed up. Although we have a relatively large sample size and patients tend to be followed up by the same surgeon postoperatively, the missing data were deleted in the analysis and we defaulted that the missing patients should not have developed cataracts. ICL V4 was introduced at our institution in 2006 and commercially launched in 2007, while ICL V4c entered the Chinese market in 2014. Consequently, the follow-up period for V4c is inherently shorter than for V4. The potential for selection bias and the underestimation of complications in the V4c group due to shorter follow-up may introduce potential for selection bias and underestimate V4c complication rates. Secondly, we used the vault at the onset of cataract in analysis because we did not have complete data from the early postoperative period to the occurrence of ASC. Thus, the correlation between dynamic changes in the vault and the development of cataracts needs further study. Thirdly, the single-surgeon design ensures internal consistency by eliminating operator-dependent variables but may limit generalizability. While surgical factors (e.g., viscoelastic removal technique) and learning curves may contribute to ICL-related complications—as evidenced by case reports linking direct irrigation in the ICL-lens interspace to transient vacuolar changes—such surgically induced complications typically manifest within 1 month postoperatively. In our cohort, the earliest cataract occurrence was at 135.77 ± 14.38 months, effectively excluding surgical etiology. Further validation will be achieved through the accumulation of more multi-center research data in the future. Additionally, it is important to note that the ultrastructural analyses are based on a small, highly selected subset and are primarily descriptive and hypothesis-generating. The molecular mechanism of ASC development remains unknown, especially in cases with ICL V4c implantation. Future studies on oxidative stress-induced ASC may help prevent cataract development after ICL.

5. Conclusions

In this large-scale cohort with a maximum follow-up of 15 years after refractive phakic intraocular lens surgery, the central hole design of the ICL V4c has significantly reduced the incidence of cataracts post-ICL surgery. Although visual acuity can often be restored after cataract extraction, complications from high myopia continue to pose challenges to vision improvement following surgery. The ASC development may be correlated with both mechanical and metabolic damage.

Study approval

The authors confirm that any aspect of the work covered in this manuscript that involved human patients or animals was conducted with the ethical approval of all relevant bodies and the study was performed in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of Eye & ENT Hospital of Fudan University (approval number: 2021018).

Author contributions

The authors confirm contribution to the paper as follows: Conception and design of study: YJ, XC and XW; Data collection: YJ, GX, and I-CL; Analysis and interpretation of results: YJ, GX, I-CL, ML, MC, BL, YL, ZM, RN; Drafting the manuscript: YJ, XC; All authors reviewed the results and approved the final version of the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (No.82171095); Project of Shanghai Science and Technology (No. 23XD1400500); Project of Shanghai Shenkang Hospital Development Center (No. SHDC12024148).

Declaration of competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Abbreviations

ACD

Anterior chamber depth

ASC

Anterior subcapsular cataract

CDVA

Corrected distance visual acuity

UDVA

Uncorrected distance visual acuity

ICL

Implantable collamer lens

IOL

Intraocular lens

LECs

Lens epithelial cells

OVD

Ophthalmic viscosurgical device

PIOLs

Phakic intraocular lenses

PC-PIOLs

Posterior chamber phakic intraocular lenses

SE

Spherical equivalent

SEM

Scanning electron microscopy

TEM

Transmission electron microscopy

WTW

White to white

Appendix A. Supplementary data

The following is the Supplementary data to this article: