Corneal astigmatism in patients having cataract surgery is common, with approximately 40% of patients having more than 1.0 diopter (D) and 10% more than 2.0 D of corneal astigmatism.1 Various techniques have been introduced to decrease corneal astigmatism at the time of cataract surgery and thus reduce postoperative spectacle dependence and maximize uncorrected distance visual acuity (UDVA). These include on-axis incisions supplemented with opposite clear corneal incisions if indicated, limbal relaxing incisions (LRIs), femtosecond laser arcuate keratotomies, toric intraocular lenses (IOLs), and refractive surgery after cataract surgery (bioptics).2–6
Limbal relaxing incisions or femtosecond arcuate keratotomies have been found to be efficacious in the management of low to moderate astigmatism (<2.5 to 3.0 D) but are less suitable for moderate to high astigmatism, which requires toric IOLs or bioptics.7,8 To our knowledge, there are no trials comparing the effectiveness of LRIs with that of femtosecond arcuate keratotomies in the management of low-to-moderate corneal astigmatism at the time of cataract surgery. The purpose of this study was to determine whether there are differences between laser-delivered and manually delivered keratotomies using vector analysis.9–12
Patients and methods
This analysis of refractive outcomes of patients treated with LRIs or femtosecond arcuate keratotomies was performed as a secondary outcome of a prospective randomized interventional case-controlled study at Guy’s and St. Thomas’ Hospital NHS Foundation Trust, London, United Kingdom. The study was approved by local Research and Development and Cambridge South Research Ethics Committee (reference 16/EE/0180). This study adhered to the tenets of the Declaration of Helsinki.
Specific to this subgroup analysis, any patient with corneal astigmatism greater than 0.9 D based on Scheimpflug tomography (Pentacam, Oculus Optikgeräte GmbH) were offered LRIs or femtosecond arcuate keratotomy as part of their cataract operation based on the initial randomization. Eyes with previous refractive or corneal surgery or corneal pathology were excluded. Partial coherence interferometry (PCI) (IOLMaster 500, Carl Zeiss Meditec AG) was performed to obtain keratometry (K) measurements for IOL formula calculation. Corneal astigmatism was measured using Scheimpflug tomography, and the measurements were used for preoperative astigmatism planning and postoperative analysis. When biometry was not possible on PCI because of the density of a cataract, A-scan ultrasound biometry (Carl Zeiss Meditec AG) was performed. All postoperative results were recorded at the 4-week follow-up.
The methods of the study are described in the publication of the main study findings.A Eyes were randomized to receive femtosecond laser–assisted cataract surgery or conventional phacoemulsification surgery. Manual limbal markings at 0 degree and 180 degrees were made on all eyes preoperatively with patients a sitting position at the slitlamp. For the markings, a needle was used scratch the corneal epithelium at the limbus; this was followed by the use of a sterile marker pen.
Femtosecond laser–assisted cataract surgery was performed using the Lensx femtosecond laser (Alcon Surgical, Inc.). The femtosecond laser was used to create the capsulotomy and fragment the lens in all patients and intrastromal femtosecond arcuate keratotomy was performed when appropriate. All cataract surgeries were performed using local anesthesia. After the femtosecond laser treatment, the patient was transferred to the operating room for the remainder of the cataract extraction. Phacoemulsification was performed using the Infiniti machine (Alcon Surgical, Inc.) in both groups. Patients having conventional phacoemulsification surgery were prepared for surgery in the same way as those in the laser arm. Instead of receiving laser pretreatment, they were brought straight to surgery and received LRIs at the start of the cataract operation. All operations were performed by experienced surgeons who had completed at least 30 femtosecond laser–assisted cataract surgery procedures (H.W.R., V.K.W., D.P.S.O.).
Limbal Relaxing Incision Group
Limbal relaxing incision parameters were calculated based on Donnenfeld’s nomogram via an online software programB based on the K readings from the Scheimpflug tomographer and the individual surgeon’s surgically induced astigmatism (SIA) values. Target induced astigmatism (TIA) was always aimed at 100% correction. Paired arcuate LRIs were always performed; when the surgeon’s preference was to operate on axis, the 2.4 mm main wound was positioned in the middle of the LRI. When anatomy or comfort dictated an off-axis approach, the surgeon’s SIA was used to modify the LRIs.
A Mendez-style ring was used to mark the steep meridians at the start of the surgery. The LRI incision was made before the commencement of phacoemulsification using a 2.4 mm keratome to incise through epithelium and Bowman layer. Next, a 600 μm guarded diamond knife was used to incise through the stroma. No corneal sutures were placed during the surgery.
Femtosecond Arcuate Keratometry Group
Femtosecond arcuate keratotomy parameters were determined by a nomogram previously reported by Day et al.7 The settings of the femtosecond laser for the arcuate intrastromal incisions were also maintained. Although this nomogram was intended to achieve up to 70% correction only, for ease of interpretation of outcome data, the TIA was defined as a 100% correction with no residual postoperative corneal astigmatism. After the femtosecond laser was docked, the horizontal meridian was manually adjusted in cases of cyclorotation.13 In cases in which either of the arcuate keratotomies overlapped with the surgeon’s planned manual wound, the main section was positioned more peripherally than the arcuate keratotomy so that it would not be involved.
Baseline characteristics were summarized for each treatment arm. Results were analyzed primarily as per intention to treat. For all evaluations of visual acuity as an outcome, patients with visually significant ocular comorbidities were excluded prospectively and those opting for a refractive target other than emmetropia were excluded from analysis of refractive outcome. Snellen visual acuities were converted to logarithm of the minimum angle of resolution notation for analysis.14 Comparative and descriptive statistical analyses included the Fisher exact test, chi-square test, and Student t tests. All statistical tests used a 2-sided P value of α equal to 0.05 unless otherwise specified. Excel software (Microsoft Corp.) was used for data entry, analysis, and graphic representation. Intraoperative or postoperative complications were defined as any event that involved unintentional trauma to an ocular structure, requiring additional treatment, or having a negative effect on the patient’s eyesight.
Analysis of corneal astigmatic outcomes based on corneal topography measurements before and after surgery was performed using the Alpins method,9–11 with calculation of the 3 following vector parameters: TIA, SIA, and difference vector. Results are presented based on the standardized graphs for reporting the outcomes of refractive surgery and IOL-based refractive surgery.12,15 Additional parameters calculated included the correction index, coefficient of adjustment, magnitude of error, angle of error, and index of success. The axis of the steep meridian was used throughout.
Four hundred twenty-seven patients were recruited to the study between August 2016 and June 2017 as per the inclusion and exclusion criteria. Twenty-seven patients withdrew from the trial before surgery. Four hundred eyes of 400 patients received surgery between November 2016 and June 2017 (200 conventional phacoemulsification surgery; 200 femtosecond laser–assisted cataract surgery).
Fifty-one eyes of 51 patients in the conventional phacoemulsification group received LRIs, of which 8 were excluded from the UDVA analysis because of visual comorbidities (6 age-related macular degeneration [AMD], 1 amblyopia, and 1 chronic central serous chorioretinopathy). Fifty-three eyes of 53 patients in the femtosecond laser–assisted cataract surgery group received femtosecond arcuate keratotomy, of which 9 were excluded for visual comorbidities (4 AMD, 2 amblyopia, 1 previous retinal detachment, 1 vitreomacular traction, and 1 central retinal vein occlusion).
Table 1 shows the patients’ demographics and preoperative values. At baseline, the corrected distance visual acuity (CDVA) was statistically significantly worse (by 12 letters) and axial length (AL) statistically significantly longer (by 0.7 mm) in the femtosecond group. Figures 1 and 2 show the TIA and SIA single-angle vector plot. The SIA was less than TIA in both groups, indicating undercorrection. However, the femtosecond arcuate keratotomy corrected more astigmatism than the LRI, as shown by the correction indices (P = .02) (Table 2). The difference vector was also lower in the femtosecond arcuate keratotomy group (P = .02), indicating better correction (Table 2 and Figures 3 and 4). Despite a greater SIA, higher correction index, and lower difference vector in the femtosecond group, there was not quite a statistically significant difference in the index of success (ratio of difference vector to TIA) between the 2 groups (P = .07). Figures 5 and 6 show the angles of error and the TIA versus SIA graphs.
Figures 7 to 10 show the 4 standard graphs for representation of refractive outcomes of cataract surgery. In both groups, nearly 60% of patients (arcuate keratotomy 25/43; LRI 24/41) attained their visual potential without requiring refractive correction (Figure 8). Eight LRI patients (20%) and 18 femtosecond arcuate keratotomy patients (44%) attained a postoperative cylinder of less than 0.50 D (P = .01) and 18 patients (44%) versus 32 patients (74%) had less than 1.00 D of cylinder (P = .003) (Figure 10). The mean corneal astigmatism was reduced from 1.38 ± 0.40 D to 0.89 ± 0.54 D in the femtosecond arcuate keratotomy group and from 1.50 ± 0.46 D to 1.17 ± 0.69 D in the LRI group (P = .02). The postoperative refractive cylinder was 0.90 ± 0.50 D and 1.18 ± 0.90 D, respectively (P = .05). The arithmetic mean of the angle of error was very small in both groups, indicating neither group had overall misalignment of treatment. However, the absolute mean indicates a misalignment of 18 to 22 degrees (P = .28) (Table 2).
Femtosecond laser treatment was delivered successfully in all cases. There was a complication relating to laser delivery in 5 cases (9.4%) (corneal abrasion in 2 cases [3.7%] and incomplete capsulotomy in 3 cases [5.6%]). The femtosecond laser was not used to create any corneal incision. In all cases, it was used for capsulotomy creation, lens fragmentation, and arcuate keratotomy creation Limbal relaxing incisions were performed in the conventional phacoemulsification surgery group in all cases. None of the femtosecond arcuate keratotomies or LRIs resulted in complications, including posterior perforation or inadvertent placement. Intraoperatively, 2 patients in the femtosecond arcuate keratotomy group sustained an anterior capsule tear and 1 patient had intraoperative floppy-iris syndrome (IFIS). Two patients in the LRI group had IFIS and 1 had iris prolapse/trauma. Postoperatively, no patient in the LRI group had complications; 2 patients in the femtosecond arcuate keratotomy group had cystoid macular edema and 1 had a steroid response leading to an intraocular pressure of 30 mm Hg at 4 weeks.
The femtosecond laser can perform, with reliability and reproducibility, several steps of cataract surgery. These include the creation of arcuate keratotomies, which are performed to reduce corneal astigmatism at the time of surgery. Although the effects of laser capsulotomy on IOL centration and refraction as well as the effects of lens fragmentation on total phacoemulsification energy have been reported, this is the first study to assess the efficacy of automated femtosecond arcuate keratotomies compared with manual LRIs during cataract surgery. Both techniques have been shown to be efficacious at reducing corneal astigmatism but have not yet been directly compared.7,16
In this study, we used the femtosecond arcuate keratotomy nomogram originally described by Day et al.,7 notwithstanding 2 important differences. First, we used a different femtosecond laser platform and second, unlike the study of Day’s group, in which the main incisions were consistently temporal, we elected to perform our main incisions on axis when possible (eg, accounting for surgical access). Using this methodology, we found that femtosecond arcuate keratotomy had a greater correction index than LRIs, indicating that the SIA was 73% of the TIA (compared with 48% for LRIs). For analysis, the TIA in the femtosecond arcuate keratotomy group was assumed to be a 100% correction; however, the nomogram for femtosecond arcuate keratotomy aims for a 70% correction to avoid too many patients being overcorrected.7 Thus, the femtosecond AK was very accurate in what it aimed to deliver. It might be assumed, therefore, that aiming for a 100% correction with femtosecond arcuate keratotomy with on-axis incisions might deliver better astigmatic correction than our results and should be the subject of further clinical studies and nomogram refinement.
Further areas for refinement include the accuracy of the femtosecond laser incisions, a better understanding of corneal biomechanics in the context of femtosecond arcuate keratotomies, and the effects of the femtosecond arcuate keratotomy on the posterior corneal curvature. A recent optical coherence tomography study of femtosecond arcuate keratotomies17 showed that the midpoint depth of the intrastromal incisions was significantly more anterior than the planned parameters and that the locations of the paired intrastromal incisions in each eye were not correlated. In studies of biomechanical properties and factors contributing to outcomes of femtosecond arcuate keratotomy18,19 the type of astigmatism (against-the-rule, with-the-rule, or oblique) were independent predictors of the efficacy of femtosecond arcuate keratotomy and corneal hysteresis had a negative correlation with the SIA at 1 to 6 months. Löffler et al.20 found that femtosecond arcuate keratotomies affected the anterior corneal curvature and total corneal refractive power, but not the posterior curvature.
In our study, there were no significant differences in the absolute or arithmetic mean angle of error. This implies that the femtosecond laser arcuate keratotomies were no better aligned than what can be achieved manually. The femtosecond arcuate keratotomy group had a significantly smaller mean difference vector (the residual correction required to achieve the TIA), and yet the index of success was not quite statistically significant between the 2 groups. The index of success is defined as the difference vector divided by the TIA, where a number closer to zero indicates greater success, and the value in the LRI group was 0.81 compared with 0.65 in the laser group (P = .07). This could therefore possibly be explained by TIA in the LRI group being 0.12 D greater.
In addition to our findings that femtosecond arcuate keratotomy had a greater correction index than LRI, we found several possible several advantages of femtosecond arcuate keratotomies over LRIs. First, they take a few seconds to program into the laser platform and for the laser platform to undertake them. In addition, although in this study we marked all eyes at the slitlamp preoperatively, several femtosecond laser–assisted cataract surgery platforms now allow integration with corneal topography and/or tomography devices and provide iris or conjunctival vessel recognition. Thus, preoperative marking of the axis is becoming redundant,13 enhancing patient and surgeon convenience. Finally, because the incisions are intrastromal, postoperative patient discomfort might be less than with LRIs and the chance of posterior or full-thickness perforation, infection, or inflammation might be lower. However, femtosecond laser technology has significant associated additional costs, and only limited additional cost and materials are required to perform arcuate keratotomies.
One key limitation of this study is that follow-up was limited to the first postoperative month and that longer term efficacy was not evaluated. The published literature reports variable results in terms of the regression of the effects of LRIs with time, although in general such corrections appear to be relatively stable after the first postoperative month.21 In a series of 263 patients by Day et al.,22 of which 87 had intrastromal arcuate keratotomies, the regression in SIA was only 0.1 D between 1 month and 6 months and was equivalent between groups that did or did not receive arcuate keratotomies. Similarly, Chan et al.6 found stability of the astigmatic correction by arcuate keratotomy between 2 months to 2 years postoperatively, and Byun et al.19 found no significant changes between 2 months and 6 months and a series of 89 eyes. This suggests that astigmatic corrections achieved at 1 month are a good indicator of efficacy, although we are following our patients at 12 months to assess the longer term efficacy.
Despite randomization, there were some statistically significant differences at baseline between the 2 groups; that is, worse visual acuity (by 12 letters) and a longer AL (by 0.7 mm) in the femtosecond group. However, there were no differences in the preoperative astigmatism or K values. We believe that the differences in acuity and AL did not play a significant role in the outcome parameters in this current study.
In summary, we found that both manual LRIs and femtosecond laser intrastromal arcuate keratotomies were safe and easy to perform, with both achieving a meaningful reduction in corneal astigmatism. However, the laser group achieved a correction of greater magnitude than the LRI cohort 4 weeks after surgery.
What Was Known
- Femtosecond laser intrastromal keratotomies and LRIs can both be used in the management of corneal cylinder at the time of cataract surgery
What This Paper Adds
- Femtosecond laser arcuate keratotomies might offer more efficacious and accurate correction of corneal cylinder than LRIs.
1. Khan MI, Muhtaseb M. Prevalence of corneal astigmatism in patients having routine cataract surgery at a teaching hospital in the United Kingdom. J Cataract Refract Surg
2. Titiyal JS, Khatik M, Sharma N, Sehra SV, Maharana PK, Ghatak U, Agarwal T, Khokhar S, Chawla B. Toric intraocular lens implantation versus astigmatic keratotomy to correct astigmatism during phacoemulsification. J Cataract Refract Surg
3. Mendicute J, Irigoyen C, Ruiz M, Illarramendi I, Ferrer-Blasco T, Montés-Micó R. Toric intraocular lens versus opposite clear corneal incisions to correct astigmatism in eyes having cataract surgery. J Cataract Refract Surg
4. Qammar A, Mullaney P. Paired opposite clear corneal incisions to correct preexisting astigmatism in cataract patients. J Cataract Refract Surg
5. Abbey A, Ide T, Kymionis GD, Yoo SH. Femtosecond laser-assisted astigmatic keratotomy in naturally occurring high astigmatism. Br J Ophthalmol
6. Chan TCY, Ng ALK, Cheng GPM, Wang Z, Woo VCP, Jhanji V. Corneal astigmatism and aberrations after combined femtosecond-assisted phacoemulsification and arcuate keratotomy: two-year results. Am J Ophthalmol
7. Day AC, Lau NM, Stevens JD. Nonpenetrating femtosecond laser intrastromal astigmatic keratotomy in eyes having cataract surgery. J Cataract Refract Surg
8. Nanavaty MA, Bedi KK, Ali S, Holmes M, Rajak S. Toric intraocular lenses versus peripheral corneal relaxing incisions for astigmatism between 0.75 and 2.5 diopters during cataract surgery. Am J Ophthalmol
9. Alpins N. Astigmatism analysis by the Alpins method. J Cataract Refract Surg
10. Alpins NA, Goggin M. Practical astigmatism analysis for refractive outcomes in cataract and refractive surgery. Surv Ophthalmol
11. Alpins NA. A new method of analyzing vectors for changes in astigmatism. J Cataract Refract Surg
12. Reinstein DZ, Archer TJ, Srinivasan S, Mamalis N, Kohnen T, Dupps WJ Jr, Randleman JB. Standard for reporting refractive outcomes of intraocular lens–based refractive surgery. [editorial] J Refract Surg. 33, 2017, p. 218-222, Available at: https://www.healio.com/ophthalmology/journals/jrs/2017-4-33-4/%7Bebecb1e6-8a50-4ded-9912-cda00a0adba2%7D/standard-for-reporting-refractive-outcomes-of-intraocular-lensbased-refractive-surgery
13. Hummel CD, Diakonis VF, Desai NR, Arana A, Weinstock RJ. Cyclorotation during femtosecond laser–assisted cataract surgery measured using iris registration. J Cataract Refract Surg
14. Lange C, Feltgen N, Junker B, Schulze-Bonsel K, Bach M. Resolving the clinical acuity categories “hand motion” and “counting fingers” using the Freiburg Visual Acuity Test (FrACT). Graefes Arch Clin Exp Ophthalmol
15. Reinstein DZ, Archer TJ, Randleman JB. JRS standard for reporting astigmatism outcomes of refractive surgery. [editorial] J Refract Surg. 30, 2014, p. 654-659, erratum 2015; 31:129. Available at: https://www.healio.com/ophthalmology/journals/jrs/2014-10-30-10/%7B8222b603-d8a1-4b64-846c-d3ddd026b1ff%7D/jrs-standard-for-reporting-astigmatism-outcomes-of-refractive-surgery#
16. Müller-Jensen K, Fischer P, Siepe U. Limbal relaxing incisions to correct astigmatism in clear corneal cataract surgery. J Refract Surg
17. Wang L, Jiang L, Hallahan K, Al-Mohtaseb ZN, Koch DD. Evaluation of femtosecond laser intrastromal incision location using optical coherence tomography. Ophthalmology
18. Day AC, Stevens JD. Predictors of femtosecond laser intrastromal astigmatic keratotomy efficacy for astigmatism management in cataract surgery. J Cataract Refract Surg
19. Byun Y-S, Kim S, Lazo MZ, Choi M-H, Kang M-J, Lee J-H, Yoo Y-S, Whang W-J, Joo C-K. Astigmatic correction by intrastromal astigmatic keratotomy during femtosecond laser–assisted cataract surgery: factors in outcomes. J Cataract Refract Surg
20. Löffler F, Böhm M, Herzog M, Petermann K, Kohnen T. Tomographic analysis of anterior and posterior and total corneal refractive power changes after femtosecond laser–assisted keratotomy. Am J Ophthalmol
21. Lim R, Borasio E, Ilari L. Long-term stability of keratometric astigmatism after limbal relaxing incisions. J Cataract Refract Surg
22. Day AC, Stevens JD. Stability of keratometric astigmatism after non-penetrating femtosecond laser intrastromal astigmatic keratotomy performed during laser cataract surgery. J Refract Surg
Dr. O’Brart is a consultant to Alcon Surgical, Inc., and Sooft Italia S.p.A. None of the other authors has a financial or proprietary interest in any material or method mentioned.
Other Cited Material
A. Roberts HW, Wagh VK, Sullivan DL, Hidzheva P, Detesan DI, Heemraz BS, Sparrow JM, O’Brart DPS. A randomised controlled trial comparing femtosecond laser–assisted cataract surgery vs. conventional phacoemulsification surgery (unpublished data)
B. Abbott Medical Optics, Inc. LRI Calculator. Available at https://www.lricalculator.com
2017, Accessed 5-6-2018