Cataract surgery with phacoemulsification of the crystalline lens and replacement with an artificial IOL in the lens capsule (in-the-bag) provides excellent visual prognosis for most patients.1 However, despite uncomplicated surgery, a spontaneous late in-the-bag IOL dislocation might occur. In regions with high prevalence of pseudoexfoliation syndrome, the reported 20-year cumulative incidence is 3.0%,2 and there is a mean interval of 6 to 10 years between cataract surgery and the IOL dislocation.3,4 There are several acknowledged surgical approaches to treat this condition, in principle, variants of either repositioning of the existing IOL or to exchange it for a new one. However, there is no clear consensus on which method is favorable, and there has been a lack of prospective studies of late in-the-bag IOL dislocation.
An important safety measure for this type of surgery is corneal endothelial cell density (ECD), which decreases with age.5 After cataract surgery, the reported endothelial cell loss (ECL) is 2.5% per year, which is up to 8 times more than that in a healthy unoperated eye.5,6 For IOL dislocation surgery, there have been concerns that IOL exchange might result in more ECL than IOL repositioning because it is considered to be a more traumatic procedure for the corneal endothelium.3,4,7
In 2013, the authors initiated a randomized clinical trial of late in-the-bag IOL dislocation to evaluate IOL repositioning with scleral sutures vs IOL exchange with a retropupillary iris-claw IOL. Preoperative characteristics, operative data, and medium-term outcomes (6 months) have been previously reported.3,8–11 A recent report included long-term visual outcomes and complications.12 The medium-term analysis revealed a borderline more pronounced ECL in the exchange group.3 The aim of this study was to examine whether there was a difference in corneal ECD between IOL exchange and repositioning of the existing IOL in a 2-year longitudinal analysis after surgery for late in-the-bag IOL dislocation. The secondary aim was to elucidate whether preoperative ECD should be a decisive factor for the surgical approach.
Between January 2013 and December 2015, patients with late-in-the-bag IOL dislocation who were suitable for both surgical techniques were invited to participate in the study. The inclusion and exclusion criteria have been published previously, and there were no restrictions on preoperative ECD.3 The study comprised 104 patients (104 eyes) randomly assigned to 2 groups: an IOL exchange group or an IOL repositioning group. The exchange surgery involved explanting and replacing the IOL and capsule for an aphakic iris-claw IOL (Verisyse, VRSA54, Abbott Laboratories, Inc.) retropupillary enclavated to the iris. Repositioning surgery involved two 10-0 Prolene sutures (Johnson & Johnson Vision Care, Inc.) through the capsular bag around each haptic and fixation to the scleral wall. In cases of vitreous loss, the surgeon either removed the vitreous strands from the incision by using cellulose sponges and scissors or, if needed, an automated vitrectomy was performed (Stellaris Vision Enhancement System, Bausch & Lomb, Inc.). A previous article describes the surgical procedures and design of the dislocated IOL in detail.3 This study adhered to the tenets of the Declaration of Helsinki; it was approved by the Regional Committees for Medical and Health Research Ethics (2012/1981) and was registered at clinicaltrials.gov (NCT01784926) before the inclusion. All patients provided written informed consent for the treatment and participation in the research project.
One surgeon (L.D.) with experience in both methods conducted all operations. Among the 86 patients who were still alive at 2 years postoperatively, 66 (77%) completed the 2-year follow-up. A flowchart and more details on patients lost to follow-up have been previously published.12 This analysis included all patients with at least 1 ECD measurement during the study period (n = 44 for IOL exchange, n = 50 for IOL repositioning).
Preoperative and Postoperative Examinations
The same ophthalmologists (M.D. and O.K.) and optometrists performed a thorough ophthalmological examination at the preoperative visit and at 6 months, 1 year, and 2 years postoperatively.
The highest preoperative intraocular pressure (IOP) between the diagnosis of late-in-the-bag IOL dislocation and the preoperative visit was included in the analysis because it was considered to be more correlated with ECL than the IOP at later postoperative visits where the mean values were within normal limits for both treatment groups.12 The ECD was measured preoperatively and at 6 months, 1 year, and 2 years postoperatively with a noncontact confocal microscope (ConfoScan 4, NIDEK Co., Ltd.). Measurements were not feasible in all cases or at every time point because of poor cooperation or technical problems.
The images were obtained from the central cornea, and by default, the program automatically counted cells within a rectangular area of 0.0935 mm2. The image used was the clearest one at each time point, and the median of 3 automatic counts was chosen for semimanual count, with the researcher masked to the group affiliation. The semimanual counting was performed by manual marking of all whole cells and partially visible cells in the upper and left frame edges, followed by the software calculation of the ECD measured in cells per square millimeter. Compared with semimanual counts, the automatic counts gave up to 30% higher values because of a software tendency to double count the cells, especially for low ECD counts. Therefore, automatic counts were disregarded, and the analysis included only the semimanual ECD counting, as also recommended by others.13 The central corneal thickness (CCT) and the anterior chamber depth (ACD) were measured with high-resolution Scheimpflug topography (Pentacam HR, Oculus Optikgeräte GmbH).
This trial used visual acuity as the primary outcome measure, and the sample size calculations required a minimum of 28 individuals in each group at the final observation.3 A substantial loss to follow-up was anticipated as a consequence of the high mean age at inclusion; hence, it was decided to include more patients. This study considered a secondary outcome from the same trial with the aim to evaluate any differences between the 2 different surgical approaches regarding the longitudinal effect on ECD.
To account for the unbalanced preoperative ECD and repeated measurements within each individual (correlated data), a linear mixed model with a random intercept was used.14,15 To evaluate whether the time effect was different between the 2 treatment groups, the model included an interaction term between group and time, together with the main effects of group and time. Furthermore, fixed effects for the following variables were considered: preoperative age, sex, preoperative dislocation grades (1–3),12 highest preoperative IOP, preoperative ACD, CCT (time-dependent covariate), complications (time-dependent covariate), and need for postoperative surgery.
The multivariate model included variables with a P value less than .2 from univariate models. Multicollinearity did not appear. A stepwise backward selection was performed where variables with a P value of more than .2 were excluded from the final model. Moreover, to account for the longitudinal design, a first-order autoregressive covariance structure for each treatment group was used. The validity of the model was checked by plots of standardized residuals vs predicted values and by normal probability plot of standardized residuals. The results presented are β coefficients with 95% CIs and P values. In addition, the longitudinal changes of ECD and CCT in the 2 treatment groups are presented graphically. Two subgroup analyses examined the 2 treatment groups for patients with low vs high ECL and patients with low preoperative ECD (<1500 cells/mm2). A third subgroup analysis examined the effect of automated anterior vitrectomy on ECL. All statistical analyses were performed using Stata Statistical Software (release 15, StataCorp LLC).
In total, 94 patients had 275 observations during the study period. There was a mean of 2.9 observations per patient (range 1 to 4). Table 1 summarizes a detailed description of the observations and preoperative characteristics.
As summarized in Table 2, the longitudinal analysis revealed a highly significant negative main effect of time, meaning that the exchange group had a significant reduction in ECD during the study period. There was also a significant difference in ECD between the 2 treatment groups preoperatively. However, there was no significant interaction between time and group, meaning that the influence of time was similar in the 2 treatment groups. Thus, the repositioning group also had a reduction in ECD during the study period.
The exchange group had an ECL of 17.5% during the 2-year period, whereas the repositioning group had an ECL of 15.3% (Figure 1). A borderline significant trend was found toward a lower ECD with increasing age. Sex, dislocation grade, highest preoperative IOP, preoperative ACD, CCT, perioperative and postoperative complications, or the need for postoperative surgery were not associated with ECL. Accordingly, the final model did not include these variables. Furthermore, the first-order autoregressive covariance structure revealed ρ estimates of 0.35 and 0.80 for the exchange and the repositioning groups, respectively.
One patient in the repositioning group had a perioperative complication (fluid misdirection syndrome), which led to an early severe ECL and corneal decompensation.12 The CCT remained stable for both groups during the 2-year follow-up (Figure 2).
The first sensitivity analysis was performed on ECL for the 71 patients with preoperative and at least 1 postoperative ECD measurement. For these patients, comparisons were made between the upper and lower quartiles of ECL relative to preoperative ECD (Table 3). Patients with high ECL (upper quartile) showed a tendency of low preoperative ECD in both treatment groups. For other covariates with a possible negative effect on ECD, the treatment groups did not show important differences.
The second sensitivity analysis investigated the possible difference in ECL between the treatment groups for patients with low preoperative ECD. The preoperative ECD was below 1500 cells/mm2 for 24 patients of which 20 patients had at least 1 postoperative ECD measurement. In this subgroup analysis, both groups had a mean follow-up of approximately 20 months. In the exchange group (n = 13), the preoperative ECD was 1213 cells/mm2 (95% CI 1095, 1331), and the ECL was −260 cells/mm2 (95% CI, −420, −99). In the repositioning group (n = 7), the preoperative ECD was 1345 cells/mm2 (95% CI, 1214, 1476), and the ECL was −428 cells/mm2 (95% CI, −639, −217).
Among patients with preoperative and at least 1 postoperative ECD measurement, automated anterior vitrectomy was used in 17 (49%) of 35 eyes in the exchange group. Because vitrector was not used in any cases in the repositioning group, the third sensitivity analysis compared eyes within the exchange group according to the use of automated anterior vitrectomy (Table 4).
To the authors’ knowledge, this randomized clinical trial is the first to perform a longitudinal analysis on corneal ECL after 2 different surgical approaches for late in-the-bag IOL dislocation. The analysis revealed similar corneal ECL in eyes undergoing IOL exchange with a retropupillary iris-claw IOLs in comparison with those treated with repositioning of the existing IOL by scleral suturing. A longitudinal study on the effect on corneal endothelial cells and corneal thickness from different types of IOLs and surgical procedures in IOL dislocation operations has been called for.16 In particular, because the ECL from surgery seems to be more pronounced for IOL exchange,3 a longitudinal study is of paramount importance because a potential long-term group difference would have implications in choosing the surgical method. Thus, this study provides important clinical information.
The 2-year longitudinal analysis showed no significant group difference and nearly linear ECL for both surgical approaches, as shown in Figure 1. The analysis found significant main effects of time and group. However, there was no combined effect from the interaction term (treatment group and time), so the observed difference between the groups reflects a preoperative difference. Thus, the groups had similar time development.
At the medium-term follow-up (after 6 months), the mixed model showed a trend toward a more pronounced ECL in the exchange group, which is in line with a previous report on 6-month follow-up data.3 This indicates a more pronounced ECL shortly after IOL exchange surgery, explained by the larger intraocular trauma with manipulation close to the corneal endothelium compared with IOL repositioning. Nevertheless, over a longer time period, this group difference seems to be offset by a similar long-term ECL. Furthermore, it seems that the cornea remains clear for more than a 2-year period with both surgical approaches and, thereby, contributes to the overall favorable visual outcome reported recently.12
Generally, intraocular surgery leads to ECL,5 and only a few studies have evaluated ECL after IOL dislocation surgery.17–20 These studies were retrospective and showed ECL varying from 11% to 21% (mean follow-up of 3 to 12 months) using different surgical approaches.17,18 Because IOL repositioning involves minimal intraocular manipulation in the anterior chamber, one could assume that this approach is less harmful for the corneal endothelium, compared with the exchange technique, where the endothelial cells could suffer from touch by instruments or by the IOL explant and implantation.7,17–19 Nevertheless, with a scleral tunnel incision and the use of a sufficient amount of ophthalmic viscosurgical device, the ECL could remain within an acceptable degree, even in exchange procedures.21 This longitudinal analysis revealed no significant difference between the groups, and the mean ECL for 2 years was 17.0% for the exchange group and 15.3% for the repositioning group. Additional surgical interventions during the follow-up period had no significant effect on ECD, which might relate to the few reoperations.12
The main analysis revealed some variation in ECL, and therefore, an explorative sensitivity analysis was performed to compare patients with high ECL (upper quartile) vs low ECL (lower quartile). For both treatment groups, low preoperative ECD tended to lead to more ECL. In addition, a sensitivity analysis was performed on patients with preoperative ECD less than 1500 cells/mm2 to investigate whether a low preoperative ECD should indicate the surgical method of choice. Because of the small samples in this sensitivity analysis, one should be careful to draw conclusions, but the data suggest that IOL exchange with a retropupillary iris-claw IOL is as safe as repositioning even when the preoperative ECD is low. The third sensitivity analysis revealed a possible association between the use of automated anterior vitrectomy and ECL within the exchange group. Probably, the use of anterior vitrectomy might reflect increased intraocular manipulation, with subsequent increased ECL.
In IOL exchange surgery, the iris-claw IOL used in this trial is the only one of several options for a new posterior chamber IOL. Alternatives include several sclerotomy techniques developed over the past decade, such as sutureless intrascleral tunnel fixation,22 fibrin glue-assisted intrascleral fixation,23 or flanged IOL fixation.24 However, from the authors' experience, the implantation of an iris-claw IOL is less of technical difficulty than other approaches. Furthermore, it is time efficient,3,10 and the implantation of the new IOL requires only minor intraocular manipulation.
The iris-claw IOL can be placed in the anterior chamber or retropupillary in the posterior chamber.25,26 Concerns have arisen regarding potential long-term differences in ECL for these 2 placements,27 and several studies have indicated that posterior chamber placement might be preferable.18,21,27,28 Nevertheless, these studies have important differences regarding indications for surgery and patient characteristics. Furthermore, the study designs have limitations to conclude about the preferred placement of an iris-claw IOL for aphakic use.
In this trial, the iris-claw IOL was enclavated retropupillary, as also preferred by others.18,21,27,28 A retropupillary placement is more physiological and preserves a deeper ACD compared with an anterior enclavation.27 It might also be favorable in cases of redislocation (disenclavation) as 1 enclavation usually still fixates the IOL and can be reenclavated relatively easily without harming the corneal endothelium.26
The Artisan (Ophtec BV) aphakia IOL is identical to the Verisyse iris-claw IOL used in this trial. It was originally made for phakic use and is not approved for aphakia in the United States. However, an ongoing clinical trial (NCT01547429) aims for an approval of aphakic use from the U.S. Food and Drug Administration. There have been concerns about its long-term effect on the corneal endothelium.29 Nevertheless, they are not necessarily applicable for the pseudophakic eye, and the iris-claw IOL for secondary implantation has increased in popularity, at least in Europe.26,28,30 For the iris-claw IOL in phakic eyes, the ACD is an important risk factor for long-term ECL.29 However, this trial shows that the ECL is not dependent on ACD in pseudophakic eyes with a retropupillary IOL.
One of the main strengths of this study is the randomized design, and to the authors’ knowledge, it is the first longitudinal analysis to evaluate the impact on the corneal endothelium from 2 different surgical approaches for late in-the-bag IOL dislocation. Analyzing repeated correlated data using a linear mixed model is preferable for several reasons. It is flexible, enables the use of all collected data, and adjusts for preoperative differences with a random intercept added to the model.14,15 Hence, the mixed model improves the validity and gives more certain estimates for the included population as a whole compared with longitudinal analysis of covariance or analysis of changes.14,15
The analysis included 94 (90.4%) of 104 randomized patients, and the ECD measurements involved semimanual counting, which is more accurate than automatic counting, particularly for low cell counts. Even though ECD measurement was not feasible for all patients at every time point, and the missing data represent a limitation, the linear mixed model yielded a mean of 2.9 ECD observations from 94 patients and comparable group sizes. Patients might have low ECD when they present with late in-the-bag IOL dislocation, and ECD studies often exclude patients with low ECD, low ACD, or glaucoma.16,19,25,29,30 This study, however, had no restrictions on preoperative ECD, glaucoma, or other ocular diseases, and the rather broad inclusion criteria improve the external validity. Yet, the results are limited to elderly patients with late in-the-bag IOL dislocation. The authors suggest, however, that the results for the exchange group might also apply to some extent for patients beyond in-the-bag IOL dislocation in need for secondary IOL implantation (eg, aphakic) where a retropupillary iris-claw IOL is chosen.
This randomized clinical trial of surgery for late in-the-bag IOL dislocation revealed that the 2-year longitudinal ECL is similar in IOL exchange with a retropupillary iris-claw IOL as in IOL repositioning with scleral sutures. Despite low preoperative ECD in quite a few participants and a considerable long-term ECL, the risk of corneal decompensation is low from a 2-year perspective. In conclusion, with the surgical techniques applied in this trial, exchanging the IOL was as safe as repositioning of longitudinal ECL.
WHAT WAS KNOWN
- A healthy endothelium is crucial for a clear and transparent cornea.
- Corneal endothelial cell density decreases by increasing age, and the yearly cell loss is higher in pseudophakic than phakic eyes.
- For late in-the-bag intraocular lens (IOL) dislocation, there have been concerns that IOL exchange surgery might give higher endothelial cell loss compared with IOL repositioning.
WHAT THIS PAPER ADDS
- Despite low preoperative corneal endothelial cell density, the long-term risk of corneal decompensation was low for the 2 surgical approaches used in this trial.
- The long-term endothelial cell loss was not higher in IOL exchange with a retropupillary iris-claw IOL compared with IOL repositioning with scleral sutures in treatment of late in-the-bag IOL dislocation.
1. Javitt JC, Brenner MH, Curbow B, Legro MW, Street DA. Outcomes of cataract surgery. Improvement in visual acuity and subjective visual function after surgery in the first, second, and both eyes. Arch Ophthalmol 1993;111:686–691
2. Mönestam E. Frequency of intraocular lens dislocation and pseudophacodonesis, 20 years after cataract surgery: a prospective study. Am J Ophthalmol 2019;198:215–222
3. Kristianslund O, Råen M, Østern AE, Drolsum L. Late in-the-bag intraocular lens dislocation: a randomized clinical trial comparing lens repositioning and lens exchange. Ophthalmology 2017;124:151–159
4. Gimbel HV, Condon GP, Kohnen T, Olson RJ, Halkiadakis I. Late in-the-bag intraocular lens dislocation: incidence, prevention, and management. J Cataract Refract Surg 2005;31:2193–2204
5. Bourne WM, McLaren JW. Clinical responses of the corneal endothelium. Exp Eye Res 2004;78:561–572
6. Bourne WM, Nelson LR, Hodge DO. Continued endothelial cell loss ten years after lens implantation. Ophthalmology 1994;101:1014–1022
7. Chan CC, Crandall AS, Ahmed II. Ab externo scleral suture loop fixation for posterior chamber intraocular lens decentration: clinical results. J Cataract Refract Surg 2006;32:121–128
8. Kristianslund O, Råen M, Østern AE, Drolsum L. Glaucoma and intraocular pressure in patients operated for late in-the-bag intraocular lens dislocation: a randomized clinical trial. Am J Ophthalmol 2017;176:219–227
9. Kristianslund O, Østern AE, Drolsum L. Astigmatism and refractive outcome after late in-the-bag intraocular lens dislocation surgery: a randomized clinical trial. Invest Ophthalmol Vis Sci 2017;58:4747–4753
10. Kristianslund O, Dalby M, Moe MC, Drolsum L. Cost-effectiveness analysis in a randomized trial of late in-the-bag intraocular lens dislocation surgery: repositioning versus exchange. Acta Ophthalmol 2019;97:771–777
11. Kristianslund O, Dalby M, Drolsum L. High intraocular pressure in eyes with late in-the-bag intraocular lens dislocation. J Cataract Refract Surg 2019;45:1043–1044
12. Dalby M, Kristianslund O, Drolsum L. Long-term outcomes after surgery for late in-the-bag intraocular lens dislocation: a randomized clinical trial. Am J Ophthalmol 2019;207:184–194
13. Kitzmann AS, Winter EJ, Nau CB, McLaren JW, Hodge DO, Bourne WM. Comparison of corneal endothelial cell images from a noncontact specular microscope and a scanning confocal microscope. Cornea 2005;24:980–984
14. Laird NM, Ware JH. Random-effects models for longitudinal data. Biometrics 1982;38:963–974
15. Twisk J, Bosman L, Hoekstra T, Rijnhart J, Welten M, Heymans M. Different ways to estimate treatment effects in randomised controlled trials. Contemp Clin Trials Commun 2018;10:80–85
16. Kwong YY, Yuen HK, Lam RF, Lee VY, Rao SK, Lam DS. Comparison of outcomes of primary scleral-fixated versus primary anterior chamber intraocular lens implantation in complicated cataract surgeries. Ophthalmology 2007;114:80–85
17. Kim KH, Kim WS. Comparison of clinical outcomes of Iris fixation and scleral fixation as treatment for intraocular lens dislocation. Am J Ophthalmol 2015;160:463–469.e1
18. Labeille E, Burillon C, Cornut PL. Pars plana vitrectomy combined with iris-claw intraocular lens implantation for lens nucleus and intraocular lens dislocation. J Cataract Refract Surg 2014;40:1488–1497
19. Eum SJ, Kim MJ, Kim HK. A comparison of clinical outcomes of dislocated intraocular lens fixation between in situ refixation and conventional exchange technique combined with vitrectomy. J Ophthalmol 2016;2016:5942687
20. Oh SY, Lee SJ, Park JM. Comparision of surgical outcomes of intraocular lens refixation and intraocular lens exchange with perfluorocarbon liquid and fibrin glue-assisted sutureless scleral fixation. Eye (Lond) 2015;29:757–763
21. Baykara M, Ozcetin H, Yilmaz S, Timuçin OB. Posterior iris fixation of the iris-claw intraocular lens implantation through a scleral tunnel incision. Am J Ophthalmol 2007;144:586–591
22. Gabor SG, Pavlidis MM. Sutureless intrascleral posterior chamber intraocular lens fixation. J Cataract Refract Surg 2007;33:1851–1854
23. Agarwal A, Kumar DA, Jacob S, Baid C, Agarwal A, Srinivasan S. Fibrin glue-assisted sutureless posterior chamber intraocular lens implantation in eyes with deficient posterior capsules. J Cataract Refract Surg 2008;34:1433–1438
24. Yamane S, Sato S, Maruyama-Inoue M, Kadonosono K. Flanged intrascleral intraocular lens fixation with double-needle technique. Ophthalmology 2017;124:1136–1142
25. Güell JL, Verdaguer P, Mateu-Figueras G, Elies D, Gris O, Amich JM, Manero F, Morral M Unilateral iris-claw intraocular lens implantation for aphakia: a paired-eye comparison. Cornea 2016;35:1326–1332
26. Forlini M, Soliman W, Bratu A, Rossini P, Cavallini GM, Forlini C. Long-term follow-up of retropupillary iris-claw intraocular lens implantation: a retrospective analysis. BMC Ophthalmol 2015;15:143
27. Gicquel JJ, Guigou S, Bejjani RA, Briat B, Ellies P, Dighiero P. Ultrasound biomicroscopy study of the Verisyse aphakic intraocular lens combined with penetrating keratoplasty in pseudophakic bullous keratopathy. J Cataract Refract Surg 2007;33:455–464
28. Hernández Martínez A, Almeida González CV. Iris-claw intraocular lens implantation: efficiency and safety according to technique. J Cataract Refract Surg 2018;44:1186–1191
29. Jonker SMR, Berendschot TTJM, Ronden AE, Saelens IEY, Bauer NJC, Nuijts RMMA. Long-term endothelial cell loss in patients with artisan myopia and artisan toric phakic intraocular lenses: 5- and 10-year results. Ophthalmology 2018;125:486–494
30. Touriño Peralba R, Lamas-Francis D, Sarandeses-Diez T, Martínez-Pérez L, Rodríguez-Ares T. Iris-claw intraocular lens for aphakia: can location influence the final outcomes? J Cataract Refract Surg 2018;44:818–826