The study included a total of 66 eyes of 37 patients who underwent classic Visian ICL implantation (Table 1). Among these, there were 4 male patients with 7 eyes, and 33 female patients with 59 eyes. The average duration of observation was 11.96 ± 0.42 years, and the mean age was 26.06 ± 0.68 years.

Before surgery, the mean UCVA (logMAR) was 1.40 ± 0.04, the mean BCVA (logMAR) was −0.04 ± 0.00, the mean SE was −10.06 ± 0.35 D, the mean IOP was 15.25 ± 0.33 mmHg, and the mean corneal ECD was 2,958 ± 40.56 cells/mm² (Table 1).

First, we evaluated long-term efficacy of ICL implantation. The mean UCVA (logMAR) showed significant improvement compared to preoperative UCVA at 1, 5, and 10 years postoperatively, with values of 0 ± 0.01, 0.04 ± 0.02, and 0.06 ± 0.02, respectively, all of which were statistically significant (mixed-effects analysis, p < 0.001) (Fig. 1A). The mean UCVA was significantly worsened (mixed-effects analysis, p < 0.05), when comparing 1 and 10 years postoperatively (Fig. 1A). However, this difference of UCVA was insignificant when cataract-occurring eyes were excluded (Fig. 1B). It indicates that improved visual acuity was well maintained after ICL implantation until the cataract developed.

The mean SE demonstrated effective myopic correction at 1 and 3 months, and 1, 3, 6, 8, and 10 years postoperatively, with values of −0.26 ± 0.06, −0.03 ± 0.07, −0.42 ± 0.07, −0.57 ± 0.06, −0.72 ± 0.09, −0.77 ± 0.12, and −0.90 ± 0.17 D, respectively, when compared to preoperative mean SE (p < 0.001). Although SE became significantly myopic from 1 month to 1 year (mixed-effects analysis, p < 0.05), there were no significant differences between 1 year and 10 years, and the values remained stable after 1 year (Fig. 1C).

Vault of ICL showed a gradual decrease over time, with values of 0.53 ± 0.04, 0.47 ± 0.04, 0.48 ± 0.03, 0.40 ± 0.02, 0.4 ± 0.02, 0.37 ± 0.02, and 0.36 ± 0.02 mm at 1 and 3 months, and 1, 3, 6, 8, and 10 years postoperatively, respectively. The vault remained stable without a significant decrease up to 1 year postoperatively, but after that period, there was a significant reduction until 10 years (p < 0.001) (Fig. 1D).

Next, we evaluated long-term safety of ICL implantation. After ICL surgery, cataract occurred in 13 eyes (19.7%) (Fig. 2A), and glaucoma occurred in 1 eye (1.5%) (Fig. 2B). There were no patients who required cataract surgery after ICL removal during the average 12-year follow-up period. The mean ECD at 1, 3, 6, 8, and 10 years postoperatively was 2,917 ± 35, 2,841 ± 38, 2,707 ± 30, 2,564 ± 35, and 2,615 ± 34 cells/mm², respectively. There was a significant decrease compared to preoperative values at 6, 8, and 10 years (mixed-effects analysis, p < 0.001) (Fig. 2C). However, a decrease to below 2,000 cells/mm² was observed in only one eye (1.5%) until last follow-up. Average ΔECL was 1.13% per year. There was no observed ECL significant enough to warrant ICL removal during an average follow-up period of 12 years. IOP measured with a noncontact tonometer at 1 month, and 1, 3, 6, 8, and 10 years postoperatively was 15.60 ± 0.30, 15.29 ± 0.29, 15.62 ± 0.34, 15.94 ± 0.35, 17.11 ± 0.35, and 16.73 ± 0.34 mmHg, respectively. Compared to preoperative values, there was a significant increase at 8 years and 10 years postoperatively (p < 0.001 and p < 0.05, respectively), indicating a gradual upward trend (Fig. 2D). Only two eyes had measurements exceeding 21 mmHg with a noncontact tonometer, but their IOP subsequently decreased to below 21 mmHg during the follow-up period. Glaucoma developed in one eye and was detected 15 years after surgery. The patient has been using topical brimonidine/timolol (Combigan, brimonidine tartrate 0.2%/timolol maleate 0.5% ophthalmic solution; Allergan Inc) and omidenepag isopropyl 0.002% (EYBELIS, Santen Pharmaceutical) to maintain IOP in the mid-teen range. In this patient, the ITC index on AS-OCT was 0%, suggesting that the occurrence of glaucoma and the previous ICL implantation surgery are unlikely to be significantly related. Among 46 eyes that could be measured for ITC at last follow-up, only 8 eyes (17.4%) showed ITC index ranging from 3.3% to 20.6%. Mean ITC index was 7.49% ± 2.10%. The eye with the highest ITC index showed 20.6% with a high vault of 0.66 mm (Fig. 3A, 3B). Although the mean vault of positive ITC group tends to be higher (0.39 ± 0.08 mm) than in no-ITC group (0.37 ± 0.02 mm), it was not significant (Fig. 3C) and did not show any correlation (Fig. 3D).

Finally, we evaluated demographic and biometric risk factors associated with either cataract or ECL. For the analysis of cataract-related risk factors, the patients were divided into those who developed cataract and those who did not. For the analysis of ECL-related risk factors, average ΔECL value (1.13% per year) was used as a reference value. Patients were divided based on whether the rate was greater than or equal to 1.13% or less than 1.13%. To determine risk factors associated with the occurrence of cataract, we analyzed age, preoperative mean SE, preoperative ECD, preoperative ACD, C-ICL, and the last measured vault value (Fig. 4A-4F). Among these factors, age and last measured vault value were statistically significant (independent t-test, p < 0.05). In the group where cataract occurred, the average age at the time of surgery was higher, at 30.08 years and the last measured vault value was on average 0.275 mm smaller. Univariate analysis revealed that age showed a statistically significant difference in relation to cataract formation (p = 0.030). When multivariate logistic regression analysis was conducted using indicators with p-values of 0.2 or lower, age still showed a statistically significant difference (p = 0.047).

When searching for risk factors associated with ΔECL of 1.13% per year or more (Fig. 5A-5F), it was found that both age and preoperative ECD were statistically significant factors (Mann-Whitney U-test, p < 0.05 and independent t-test, p < 0.001, respectively). In the group with ΔECL of 1.13% per year or more, the average age at the time of surgery was lower, at 24.28 years and the preoperative ECD was higher, with an average of 3,104 cells/mm². In univariate analysis, preoperative ECD showed a statistically significant difference in relation to ΔECL (p = 0.007). When multivariate logistic regression analysis was conducted using indicators with p-values of 0.1 or lower, preoperative ECD still showed a statistically significant difference (p = 0.014).

When we analyzed the correlation between ΔECL and either preoperative ACD or C-ICL, no significant correlation was found (Supplementary Fig. 2A, 2B). Additionally, there was no correlation between myopic shift and either vault or lens thickness (Supplementary Fig. 3A-3B).

This study is worthy of notice reporting the safety and effectiveness of posterior phakic IOL implantation surgery for high myopia correction for a mean follow-up of 12 years and providing quantitative biometric evaluations to allow an objective analysis of risk factors. Additionally, this study suggests that regular monitoring should be required beyond 12 years to assess cataract, ECL, and IOP.

Although many studies show better efficiency of posterior phakic IOL for moderate to high myopic correction and some studies present long-term favorable outcomes of posterior phakic IOL [3-8,10,11], we are still concerned about adverse events such as long-term visual decrease, myopic regression, cataract, ECL, and glaucoma despite their low incidences [5,12-19]. That is why we conducted this study to reveal those outcomes and related risk factors.

As for long-term visual acuity changes, many studies have reported mild myopia regression after ICL implantation surgery [18,20], but the exact causes have not been fully elucidated. Factors such as changes in axial length before and after surgery [18,20], vault reduction [21], and cataract formation [18] have been suggested. At first, we too suspected vault reduction as a plausible cause for myopic regression (-0.64 D per 10 years) because vault height settled down continuously after 1 year (Fig. 1D). However, correlation analyses between vault profiles and myopic shift did not show significant relationships (Supplementary Fig. 3A, 3B). Further large-scale research is necessary to clarify the relationship between vault and myopic shift. When we looked at the relation between the lens thickness at last follow-up and the delta SE, there was no correlation between lens thickness and myopic shift (Supplementary Fig. 3C) and lens thickness in cataract group was not different from that in noncataract group (data not shown). Therefore, myopic shift is also unlikely to be related with lens thickness. Considering the unrelation with either vault or lens thickness to the myopic shift, there is another possibility that hydrophilic material itself would affect the changes of refractive diopter in ICL.

When comparing 1 year and 10 years postoperatively, a significant decrease of visual acuity was observed. Given that this difference disappeared after excluding the cataract occurrence group, cataract may be a primary cause of decreased visual acuity.

The development of cataract after ICL implantation surgery is a significant and possible complication. Various factors have been suggested as potential causes, including mechanical damage to the crystalline lens during ICL insertion, contact between the ICL and the crystalline lens after surgery, and chronic inflammation in the anterior chamber [22,23]. In our study, conducted with an average follow-up of 12 years, cataract occurred in 13 cases (19.7%). In other domestic and international studies with over 10 years of follow-up, the incidence rate of cataract ranged from 6.2% to 54.8% [4,5,8].

As for risk factors, the average age at the time of surgery was higher in the cataract group, at 30.08 years, and the last measured vault during the observation period was smaller, with an average of 0.275 mm. Previous studies have also suggested a correlation between low vault and cataract development, which corresponded with the outcome of this study [5,21,24]. Age is a known risk factor associated with cataract development [25]. Lee et al. [26] mentioned that as individuals age, the thickness of the crystalline lens increases, and as the crystalline lens thickness increases, there is a higher likelihood of reduced vault, potentially increasing the risk of cataract development.

Although it is rare, elevated IOP and the development of glaucoma are also possible complications after ICL implantation surgery. Complications such as angle-closure glaucoma and pigment dispersion syndrome can occur, potentially requiring ICL removal [27]. In this study, IOP remained within the normal range, although it showed a significant increase compared to preoperative values. There were only two eyes that had measurements exceeding 21 mmHg, and both returned to normal range during the follow-up. Glaucoma occurred in one eye (1.5%) which was well controlled with antiglaucoma drops. These results demonstrated that the occurrence of glaucoma was not a major problem during the 12 year-follow-up period after

ICL implantation surgery. Considering ITC index, anatomical change by ICL implantation surgery might not be significantly related with glaucoma and further research is necessary to investigate other contributing factors. ICL is known to have less impact on the corneal ECD loss compared to anterior chamber IOLs [28,29]. Nevertheless, several studies have mentioned changes in corneal ECD loss after ICL implantation surgery [7,18,30]. In this study, the mean corneal ECD showed a significant decrease for 6 years postoperatively compared to preoperative values. However, a decrease to below 2,000 cells/mm² was observed in only one eye (1.5%). This suggests that ECL is not clinically relevant until 10 years. In this study, the average ΔECL was 1.13% per year, which is approximately twice the rate of loss attributed to normal aging, estimated at 0.5%. Previous research has shown a wide spectrum of ΔECL, ranging from 0.53% to 4.4% per year [6,18,30]. Factors contributing to such variability may include the basic characteristics of the study population (such as race and gender), differences in sample size and duration of follow-up, as well as the proficiency of the surgeons involved [31]. Yang et al. [32] mentioned that an average follow-up period of 52 months was observed, when comparing the ECD values between the last follow-up visit and preoperative measurements, it revealed an average decrease of 4.03%, and it was noted that vault was the most significant factor associated with ECD change. Yoon et al. [33] also mentioned that high vaulting is a significant factor in ECL among patients who have undergone ICL. Unlike the previous study [33], our study did not show the relation between the ECL and vault, C-ICL, or preoperative ACD. We believe different preoperative ACD may affect the outcome of our study profoundly. The mean preoperative ACD was 2.9 mm in the previous study [33], whereas mean preoperative ACD was 3.35 mm in our study, which is considerably deeper than the preoperative ACD in the previous study (a proportion of the eyes with preoperative ACD less than 2.9 mm was only 9% in our study). Also, the mean distance of C-ICL was 2.98 mm in our study. The differences between our study and previous study were low ΔECL (about 1.1% per year) and no correlation between preoperative ACD and ECL. We hypothesize that the risk factors for ECL caused by insertion of ICL relates to two ways: (1) subclinical inflammation due to ICL per se or posterior iris rubbing by contact of ICL and (2) anatomical angle narrowing to affect ECL. Our study suggests that the effect of subclinical inflammation is the major factor on ECL not an anatomical angle narrowing in eyes with mean preoperative ACD deeper than 3.3 mm, resulting in 2.98 mm of mean C-ICL. Hence, the analysis that preoperative ACD as well as C-ICL has no relation with ΔECL in our study. However, based on previous study, it seems that ECL can be affected by both two reasons (subclinical inflammation and narrow angle) in eyes with preoperative ACD that was about 2.9 mm or less, resulting in more loss of ECD and showing significant relation between preoperative ACD and ECL. Although ΔECL was low as 1.1% per year, it can reach 22% of ECL by 20 years after the surgery. Taken together, annual monitoring of ECL is required until the ICL removal. In the analysis of risk factors for ΔECL, it was observed that the group demonstrating a higher ΔECL exhibited a higher preoperative ECD, with a mean of 3,104 cells/mm², and a younger age at the time of surgery, with a mean age of 24.28 years. It can be speculated that, generally, younger individuals would have a higher preoperative ECD, and the combination of younger age and higher preoperative ECD may be classified as risk factors for a higher ΔECL. Also, it suggests that corneal endothelial cells in younger individuals may be affected more than in older individuals with intraocular subclinical inflammation. This study adhered to the universal age criteria of 18 years and above. However, there was no clear-cut criteria regarding the minimum age required for phakic IOL implantation surgery. Base on this study data, when determining the timing of surgery in consideration with endothelial cell stability, patients under 25 years old require thorough consultation before proceeding with posterior phakic IOL implantation.

The limitation of this study is that it is retrospective in nature and was conducted on patients with an average follow-up of 12 years and that the sample size is small. For a more accurate analysis of complication occurrence, prospective long-term studies may be necessary in the future.

In conclusion, this study demonstrated the high safety and efficacy of ICL implantation surgery until an average follow-up period of 12 years. The average ΔECL (1.13% per year) in individuals who underwent surgery was higher than individuals who did not undergo surgery, which is 0.5% per year. This highlights the need for regular follow-up observations for ECL as well as cataract. While IOP remained within the normal range, there was a trend of gradual elevation, emphasizing the importance of more extended and meticulous long-term monitoring. Age was identified as a factor related to the annual rate of corneal ECD loss and the occurrence of cataract.