This study compared the long-term outcomes of the LASIK and phakic IOL implantation in the management of post-PK refractive errors. Several studies have reported the outcomes of LASIK and phakic IOL implantation after keratoplasty [6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24]. However, none of these studies compared the outcomes of these two procedures. In our practice, Artisan IOL is offered to post-PK patients when LASIK is not indicated such as high levels of myopia and inadequate corneal graft thickness. This explains the statistically significant difference in preoperative UDVA and spherical equivalent refraction between the groups. We included patients who had chiefly regular symmetric astigmatism after corneal transplantation, based on preoperative CDVA and corneal tomography. The lower CDVA in Artisan group could be due to minification effect of spectacles correction for larger myopic refractive error. LASIK can correct cylindrical component of refraction up to 6 D and non-toric Artisan phakic lenses have only spherical powers. Relaxing incision with concomitant counterquadrant compression sutures were performed to reduce preoperative astigmatism to ≤6 D in the LASIK group and to ≤4 D in the Artisan group, which explains the significant difference in baseline astigmatism between two groups.

The results of our study demonstrate no significant change in UDVA after LASIK. None of the eyes in this group had UDVA ≥20/40 postoperatively. However, UDVA significantly improved after Artisan implantation and 62.5% of the eyes had UDVA of ≥20/40 at the final visit. In addition, the Artisan group had a lower postoperative refractive error compared to LASIK despite higher preoperative refractive error; 6.3% versus 58.8% of eyes were within ±2 D of emmetropia after LASIK and Artisan implantation, respectively. Postoperative CDVA significantly decreased in the LASIK group, whereas it remained stable in the Artisan group at the final follow-up examination compared to preoperative values. Artisan group had a greater gain of CDVA, probably due to retinal image magnification after correcting a high level of myopia. Loss of CDVA in the LASIK group could be due to irregular astigmatism caused by surgery.

Literature review suggests less favorable outcomes with LASIK in our study. Reported UDVA ≥20/40 and CDVA ≥20/40 after LASIK in post-PK eyes varies from 28 to 86% and 74 to 100%, respectively [6,7,8,9,10,11,12,13,14,15,16,17], in contrast to 0 and 87.5% in our study. Similarly, the loss of 2 lines of CDVA in 18.8% of the LASIK-treated eyes in our series is higher than that reported by other studies (0 to 16%) [7,8,9,10,11,12,13,14,15,16,17]. An improvement of 2-line or greater of CDVA was not observed in our LASIK eyes, in contrast to the reported range of 5 to 56% [6, 8]. The reported reduction in spherical equivalent refraction ranges from 74.8 to 92.9% and in cylinder is 42 to 87.9% in the postkeratoplasty LASIK series [9, 10, 19]. We observed a 40% reduction in spherical equivalent refraction and no change in astigmatism after LASIK. However, most of the LASIK series for postkeratoplasty refractive error had either a short follow-up duration or included PK for variety of etiologies. Our short-term outcomes are in line with the reported efficacy and safety of the LASIK procedure in patients with PK [6,7,8,9,10,11,12,13,14,15,16,17]. Optimal refractive correction was achieved as early as month-3 post-LASIK in our study, however, our results demonstrated a regression of its effect 3 years postoperatively. There are 3 studies on the long-term outcomes of LASIK for post-PK refractive errors in keratoconus [8, 15, 20]. Kwitko et al. [8] included 14 patients of which 13 had keratoconus, and followed the patients from 12 to 42 months after LASIK. The authors reported 2 lines or more of improvement in the UDVA and CDVA in 64.3 and 21.4% of patients, respectively [8]. Mean reduction in refractive astigmatism was 47.5% with no astigmatism reduction in 28.6% of cases, and progressive changes in refraction were seen in 35.7% postoperatively [8]. In addition, retreatment was necessary in 42.9% of cases in their series [8]. In another study on the 5-year outcomes of LASIK in 30 eyes that had undergone PK for keratoconus, spherical equivalent refraction decreased from − 7.15 D preoperatively to − 0.97 D at postoperative year-1 and -1.05 D at postoperative year-5 [15]. At the last visit, 53.3 and 86.7% of eyes had refractive astigmatism ≤ − 1 D and ≤ − 4 D, respectively [15]. In a similar study, Spadea et al. [20] performed LASIK using standard technique (15 eyes) versus topographically guided two-stage technique (15 eyes) and followed the patients for 36 months. Visual acuity and refraction were significantly improved in both groups, with better outcomes achieved with the two-stage approach [20]. Their visual and refractive outcomes were stable through the course of the study [20]. Progressive changes observed in the long-term follow-up in our study could be due to including recipient corneas within a lamellar flap that was larger than grafts. Lamellar cut of the recipient corneal tissue with the typical altered biomechanics in keratoconus could result in progressive myopia and astigmatism from peripheral corneal ectasia. Another explanation for significant change in post-LASIK refraction could be the small optical zone (5 mm) which could result in regression of astigmatism correction postoperatively.

Our favorable outcomes with phakic Artisan IOL are in line with previous reports. Tahzib et al. [22] implanted Artisan toric phakic IOL in 36 PK eyes for various corneal pathologies including keratoconus (13.6%). After a mean follow up of 28.5 months, spherical equivalent refraction was decreased from − 3.19 ± 4.31 D to − 1.03 ± 1.20 D (77.8% reduction) and refractive astigmatism from − 7.06 ± 2.01 D to − 2.00 ± 1.53 D (88.8% reduction). Postoperative UDVA and CDVA ≥20/40 were achieved in 31.6 and 80.6% of their patients, respectively, and they observed a loss of CDVA of > 2 lines in 8.3% of eyes, and a gain of ≥2 lines in 8.3% [22]. Alfonso et al. [23] implanted posterior chamber spherical (for astigmatism < 2.5 D) or toric (for astigmatism > 2.5 D) implantable Collamer lens (ICL) in 15 PK eyes. At the postoperative month 24, spherical equivalent refraction was decreased from − 9.80 ± 5.88 D to − 0.95 ± 1.12 D. Postoperative UDVA and CDVA ≥20/40 were achieved in 46.6 and 80% of their patients, respectively. No eye lost 1 or more lines and 46% gained 1 or more lines of CDVA, whereas 52% had no change in their CDVA [23]. Iovieno et al. [24] reported the outcome of toric ICL implantation in 7 keratoconus eyes with previous history of PK (n = 6) or deep anterior lamellar keratoplasty (n = 1). After a mean follow up of 12 months, toric ICL was reported to be very predictable in correcting the spherical refractive error, with a mean postoperative spherical refraction of 0.53 ± 0.75 D. In addition, good result was achieved in astigmatism correction as 71.4% of the eyes had postoperative refractive astigmatism ≤1.5 D. The percentage of eyes with UDVA and CDVA ≥20/40 was 87.5 and 100% in their study [24]. Tahzib et al. [22] and Alfonso et al. [23] observed no regression in spherical equivalent refraction and astigmatism after phakic IOL implantation. Similarly, the stability of the postoperative spherical equivalent and refractive astigmatism was excellent after Artisan lens implantation in our study, yet vector analysis revealed a shift in the axis of the postoperative astigmatism compared to the preoperative value. This shift was probably due to surgically induced astigmatism by placing a 6-mm incision to insert the rigid phakic lens.

The possibility of accelerated endothelial cell attrition of corneal graft is a concern with phakic IOL implantation after keratoplasty. In their series, Tahzib et al. [22] observed 34.8% reduction in endothelial cell count 4 years after Artisan IOL implantation and 3 eyes developed graft failure due to irreversible endothelial rejection (n = 2) and progressive endothelial cell loss (n = 1) [22]. They included eyes with preoperative endothelial cell count as low as 500 cells/mm2. However, we did not observe any graft failure or significant loss of endothelial cells as an endothelial cell count of > 1500 cells/mm2 was required prior to phakic IOL implantation. Similarly, Alfonso et al. [23] implanted ICL in eyes with an average preoperative graft endothelial cell density of 1660 ± 427 cell/mm2 which was reduced to 1526 ± 398 cell/mm2 at 24 months postoperatively (a reduction of 8.1%) [23]. All grafts remained clear in their study at the final visit [23].

There are several limitations to our study. First, we used a spherical phakic IOL because we had no access to the toric model during the study period. Relatively better correction of the astigmatism in the previous similar studies is due to implanting toric IOL [22,23,24]. Second, we did not evaluate the effect of LASIK procedure on the corneal graft endothelium, and postoperative endothelial cell count was measured 1 year after Artisan implantation; therefore, long-term safety of these procedures with respect to corneal graft endothelium cannot be evaluated by our results. Third, topography was not performed postoperatively to evaluate decentered ablation, irregular astigmatism, or even the recurrence of keratoconus in the LASIK group. However, decentered ablation could not be responsible for increase in postoperative astigmatism in our LASIK group as all patients attained good refractive outcomes up to postoperative month 36. The small sample size could be another limitation of the current study. The result of post hoc power analysis, however, revealed that our study has a power of 95.9% to detect the observed difference in final spherical equivalent refraction between two groups.