The introduction of intraocular lenses (IOLs) in 1949 has been transformational, helping the vast majority of patients achieve better optical quality of vision and an enhanced quality of life.1,2 However, there are occasions when the IOL becomes a detriment and requires removal. Although the rate of removal per cataract surgery is probably decreasing, the total number of IOL removals and exchanges seem to be increasing.3–5 This circumstance might be categorized into early and late complications. Previous studies of IOL removal have largely consisted of data from anterior segment surgeons, commonly for refractive reasons, glare, or early dislocations, which likely account for the majority of early removal.3 IOL removal, especially at a later date and characteristically by vitreoretinal surgeons, is pursued to treat IOL dislocation, uveitis or glaucoma, opacities on the IOL (more commonly representing an impediment to executing posterior segment therapy as opposed to being primarily visually detrimental), and IOL breakage.6 The need to remove the IOL may not have been apparent prior to the vitreoretinal surgery and, hence, may become an intraoperative decision. The structural design, preparation, or material of the IOL may factor into the reason for the IOL removal, or even the technique of removal.
Studies of the setting and management of IOL removal from the vitreoretinal perspective are scant, mostly in the context of a management option for dislocation of the IOL.7 Removal of an anterior chamber (AC) IOL may be indicated largely for its own reasons and has been described fairly widely.8 Posterior chamber (PC) IOL removal may become an important, or even essential, step to allow visualization (or even diagnosis) of posterior segment disease, typically, in the repair of a complex, recurrent retinal detachment. An important element to maximize success during any vitreoretinal surgery is optimal visualization. The IOL may compromised visualization in a variety of ways: opacification of the IOL, dislocation, or damage, or an IOL might impede treatment, such as in Propionibacterium acnes endophthalmitis. In such cases, it might be advisable to remove the IOL.
The purpose of this study is to describe the indications and outcomes of IOL removal encountered in a vitreoretinal surgery practice.
The study cohort included 63 eyes of 63 patients (30 males and 33 females). There were 28 (44%) right eyes and 35 (56%) left eyes. The median interval from initial cataract surgery to IOL removal was 998 days, (range: 37 days to 5,137 days). The median follow-up interval after IOL removal was 302 days (range: 45 days to 2,834 days). The type of previous cataract/IOL surgery was phacoemulsification in 48 eyes (76%), pars plana lensectomy in seven eyes (11%), extracapsular cataract extraction in five eyes (8%), and unspecified in three eyes (5%). The primary procedure was performed at different facility in 54 eyes (86%) and at the current study facility in nine eyes (14%).
Complications after the primary cataract procedure that resulted in subsequent vitreoretinal surgery included retinal detachment (RD) (n = 16; 25%), recurrent RD (n = 18; 29%), dislocated or decentered IOL (n = 14; 22%) posterior capsular rupture (n = 10; 16%), endophthalmitis (n = 8; 13%), epiretinal membrane (n = 8; 13%), vitreous hemorrhage (n = 5; 8%), proliferative vitreoretinopathy (PVR) (n = 4; 6%) retained lens fragment (n = 2; 3%), choroidal detachment (n = 3; 5%), pseudophakic bullous keratopathy (n = 2; 3%), hyphema (n = 1; 2%), chronic uveitis (n = 1; 2%), and IOL exchange (n = 1; 2%).
There were 12 patients (19%) with a documented history of Nd:YAG laser capsulotomy. The mean (standard deviation [SD]) logMAR BCVA prior to removal of the IOL was 1.98 (0.72) (Snellen equivalent 20/1,909). The mean (SD) logMAR BCVA of the fellow eye prior to removal of the IOL was 0.56 (0.86) (Snellen equivalent 20/72).
The composition of the IOL was silicone (22 eyes, 35%), unspecified (17 eyes, 27%), acrylic (14 eyes, 22%), or polymethyl methacrylate (PMMA) (10 eyes, 16%). The removed IOL was a PC-IOL in 59 eyes (94%), and an AC-IOL in four eyes (6%). The removed IOL was toric lens in five eyes (8%), and a multifocal lens in one eye (2%).
Reason for Removal
The most common reason for removal was IOL opacities (n = 42; 67%), including nonspecific opacities (n = 19; 45%), silicone oil artifact (n = 17; 40%), opaque, avascular membranes (n = 4; 10%), and fibrovascular proliferation (n = 2; 5%). The second leading indication for removal was IOL dislocation (n = 14; 22%), followed by endophthalmitis related (n = 7; 11%), and broken haptic (n = 1; 2%). There were 35 (56%) total IOLs that were removed due to obstruction in visualization during concurrent vitreoretinal surgery; these included 32 (91%) cases of IOL opacification and three (9%) dislocated IOLs (Table 1).
Reason for IOL Removal
Among the 42 IOLs removed due to opacification, 19 (45%) were silicone lenses, 14 (33%) were unspecified, five (12%) were PMMA, and four (10%) were acrylic. The 17 IOLs removed due to oil opacification included 15 (88%) silicone and two (12%) that were unspecified.
IOL removal was performed in conjunction with vitrectomy in all eyes; 51 (81%) were left aphakic, and 12 (19%) had an IOL placed in the sulcus. No AC-IOLs were implanted in this setting.
Postoperative complications included recurrent retinal detachment (n = 13; 21%), hypotony (n = 8; 13%), phthisis bulbi (n = 8; 13%), corneal edema (n = 7; 11%), cystoid macular edema (n = 5; 8%), elevated intraocular pressure (n = 3; 5%), vitreous hemorrhage (n = 3; 5%), hyphema (n = 1; 2%), anterior synechiae (n = 1; 2%), and subretinal hemorrhage (n = 1; 2%).
The mean (SD) immediate, 3 months, and final BCVA in logMAR were 2.18 (0.47), 1.85 (0.82), and 1.97 (0.85). The Snellen equivalents were 20/3,027, 20/1,415, 20/1,866, respectively. The final average (SD) BCVA in the fellow eye was 0.60 (0.89) (Snellen equivalent 20/80).
The current study describes a vitreoretinal experience of the indications and settings in which IOL removal was performed, finding that the most common indication was to remove it as a visual impediment to management of posterior segment abnormalities. The eye that requires IOL removal is characteristically a complex eye, and this likely frames the poor visual results related in this study — recurrent retinal detachments, hypotony, or anterior segment complexities.
IOL removal has been a rare necessity ever since the seminal discovery and rapid rise in IOL implantation. Studies in the late 1980s and early 1990s, commonly from anterior segment surgeons, delineated indications, initially for AC-IOLs, such as bullous keratopathy, uveitis-glaucoma-hyphema syndrome and cystoid macular edema.10–13 Later, with the advent of PC-IOLs, again predominantly the anterior segment literature, described the variety of other reasons why an IOL might be exchanged or removed.
Reports of IOL removal in the past two decades have most commonly reported it as a method for treating PC-IOL dislocation, decentration, and refractive surprises.3,8,10,14,15 A 257-subject series reported indications for removal of mostly PC-IOLs: dislocation (56.3%), incorrect lens power (19.0%), IOL opacification (11.3%), pseudophakic bullous keratopathy (2.3%), and endophthalmitis (1.9%).16 IOL removal with exchange, is one of the more common anterior segment surgeon's indications for removal, especially with the higher standard of postoperative refractive expectations and the advent of multifocal PC-IOL. A study from a tertiary referral center reported 18.3% of IOL exchanges involving multifocal IOLs in patients with asthenopia, glare, halos, and visual distortion.17
The vitreoretinal surgeon is generally confronted with IOL removal in a very different setting. These include late dislocation7,18 or addressing the IOL as an impediment to repairing a retinal detachment.19 The current study, the largest to review the indications and results of IOL removal in an exclusively vitreoretinal setting, found these as the two most common reasons for IOL explanation with IOL opacities as the leading reason (67% eyes). This is in contrast to other, predominantly anterior segment studies which report IOL opacification as the indication in between 1.3% and 29%.4,16 IOL opacification may lead to an impaired BCVA in some patients, but more commonly presents to the vitreoretinal surgeon as an optical impediment to optimal visualization for effecting retinal reattachment. Opacities such as hematologic or inflammatory debris, or limited fibrotic membranes can often be aspirated or excised without IOL removal, but this is not always the case with more exuberant opacities. Thus, the current study highlights IOL removal as an option that might offer the vitreoretinal surgeon who is struggling for an adequate view a transformational option in a given case. It is common that the decision to remove the IOL arises intraoperatively, rather than as a planned component, so it might be valuable to have pondered this possibility (and its instrumentation requirements) prior to confronting the situation. On a more sobering note, the poor visual results are a testament to the complexities that constrain vision in such eyes.
The incision for removing a foldable IOL (the dominant genre currently) generally has to be larger than for inserting. For late IOL dislocations, there is a variety of surgical methods reported for IOL removal including the enlargement of a sclerotomy, creation of a limbal wound,7,18 or segmentation of the IOL. Although not a subject of investigation in this series, it is our preference to position the IOL into the anterior chamber, and then make a conjunctival incision and create a scleral tunnel through which to remove the IOL. Avoiding a clear corneal incision and using viscoelastic may dampen what is commonly already a moderate degree of coexisting corneal decompensation. The scleral incision also lessens the potential for ongoing wound leakage of the coexisting inflamed tissues, and the visualization challenges that presents for subsequent completion of the vitreoretinal surgery. Still, the prognosis is already poor when this maneuver becomes necessary.
The second most common indication for IOL removal was still invoked as a method to treat a dislocated IOL. Since IOL repositioning is more commonly pursued in that setting by these authors, that bias likely skewed the distribution of indications. There are a multiplicity of IOL repositioning techniques advocated, with excellent results in most series. The authors prefer using scleral suture fixation, but good results have been reported with various sutureless repositioning using a variety of scleral tunnel fixation techniques.20 The sutureless techniques are really only suited for three-piece IOLs, however.
Despite the versatility of scleral suturing techniques in our experience, some IOLs cannot be satisfactorily repositioned. The common reasons for managing IOL dislocation by removing (with or without exchanging) depend upon the style of the IOL and the extent of associated ocular pathology such as a concurrent retinal detachment or anterior segment conditions. These constraints are often discernable preoperatively but must be anticipated as a possibility even when repositioning is planned. Hence, it is important to have measured for an IOL (or ascertain the power of the existing IOL) so that if exchange is pursued, the IOL power can be optimally selected. If a scleral sutured IOL is to be inserted, the calculations (characteristically made for bag fixation) should adjust for the fact that the IOL will more closely approximate sulcus fixation; the more anterior position of the IOL requires a reduction of about a diopter in the IOL power selected to avoid leaving the eye with an excessively myopic result. Such an adjustment is not necessary for an AC-IOL.
The choice of an AC- versus a PC-IOL is probably more a matter of surgeon preference. The replacement IOL may be inserted during the same surgery by inserting an AC-IOL or a PC-IOL (typically scleral fixated) In one study of in-the-bag dislocations from an anterior segment setting, the IOL was exchanged in 72% of cases, repositioned in 18%, and removed without exchange in 9.8%.21 Only 19% of cases in the current series represented management in this way.
The authors prefer to scleral suture the IOL figuring it lessens the loss of corneal endothelial cells, but this practice has not been proven, and either choice would seem suitable. To optimize centration and minimize torsion, care is given to placing the sutures about 1 mm posterior to the limbus; a more posterior location directs the haptic posterior to the sulcus and induces torsion and a more anterior placement may lead to iris chafing. Another important step is to ensure that the sutures are introduced exactly 180 degrees from each other to optimize centration. The authors use 9-0 prolene, which seems to minimize the late dislocation, a problem reported with techniques that used 10-0 prolene.22 The Gore-Tex suturing technique has been reported with good outcomes and is another option to avoid this and might be a better choice in the pediatric patient due to the higher tensile strength of the suture.23 Using scleral flaps and leaving the sutures longer (about 3 mm) externally allows them to lay flatter and results in less erosion through the conjunctiva.
A less common indication for IOL removal in this series was the optical impediment to the patient, in the absence of retinal detachment. These represent opacities or optical aberrations secondary to interaction of silicone oil with the IOL and are almost exclusively seen with silicone IOLs, as was the case in the current series. Experimental studies have shown that other commonly used materials are not immune to silicone oil opacification.24 PMMA, but not acrylic lenses, were encountered with opacification in the current study experience; snowflake degeneration was not encountered. Frank opacification of the IOL is a much less common finding and can be due to crystallization with viscoelastics, staining with intraoperative dye, and hydration of silicone lenses. There have also been reports of IOL opacification in conjunction with certain systemic medications.25
A serious limitation of this study is its retrospective in design; hence, there was no capability to make comparative BCVA comparisons. Also, there were no histopathologic analyses performed that might have further characterized debris on the IOL. As has been mentions, this was a study biased in that it represents distributions of IOL explantation that present to a vitreoretinal surgeon would encounter, and that is further biased by the authors' inclination to reposition rather than exchange otherwise standard dislocated IOL cases.
In conclusion, the settings in which a vitreoretinal surgeon might find IOL removal have been tabulated. The option of IOL removal will vary depending upon clinical context, but can be an important option when obscuring the view to the posterior pole preventing retinal detachment repair. The poor prognosis with IOL removal is probably most associated with the comorbidities of the eye and the indication for removal.