Iatrogenic Keratoconus: Corneal Ectasia Following Laser In Situ Keratomileusis for Myopia FREE

Laser in situ keratomileusis (LASIK) to correct myopia is performed by partially resecting a prescribed thickness of stroma, removing corneal tissue from the exposed stromal bed using the excimer laser, and then replacing the resected stromal tissue. This results in a substantial reduction of the biomechanically effective stress-bearing thickness of cornea provided by the residual stromal bed. There is concern that at some point, the tensile strength of the cornea might be reduced to the degree that progressive ectasia ensues, thereby resulting in steepening of the cornea, irregular astigmatism, and progressive myopia. This becomes a particularly contentious issue when, in the absence of classic clinical evidence of keratoconus, inferior steepening of the cornea seen on corneal topographic scan suggests the possibility of subclinical keratoconus. Herein, we report such a case of progressive ectasia following LASIK.

A 23-year-old Hispanic man sought refractive surgery to correct myopia. Prior to surgery, he relied on spectacles to correct his vision. He reported infrequent changes in his prescription and good visual acuity in the years prior to consultation. His optometrist's records prior to surgery documented a refraction of –12.75 – 2.25 × 65 OD and –8.50 – 1.50 × 79 OS, yielding a visual acuity of 20/30 OU. Findings from slitlamp examination (N.C.C.) revealed no characteristic corneal findings of keratoconus in either eye, including Vogt's striae or a Fleischer ring. Ultrasound pachymetry measurements prior to surgery were 555 µm OD and 560 µm OS.

Corneal topographic scans of the right eye performed prior to surgery demonstrated mild inferonasal steepening with a maximum power of 44.5 diopters (D), simulated keratometry readings of 43.0 × 134/41.5 × 44, and a minimum keratometry reading of 41.3 × 37 (Figure 1, top). Keratoconus screening by Rabinowitz/McDonnell criteria suggested the presence of keratoconus, while screening by Klyce/Maeda criteria identified a 15% similarity to keratoconus. Corneal topographic scans of the left eye showed a homogeneously regular central corneal contour with a central corneal power of approximately 41.28 D (Figure 1, bottom). Keratoconus screening by both Rabinowitz/McDonnell and Klyce/Maeda criteria failed to detect any similarity to keratoconus.

The patient underwent bilateral simultaneous LASIK (N.C.C.). A 130-µm flap was created using a manual microkeratome, and ablations 165 µm and 140 µm in depth were performed using an excimer laser (Summit Apex Plus; Summit Technologies, Waltham, Mass) in the stromal beds of the right and left eyes, respectively, estimated to leave residual stromal beds of 260 µm OD and 290 µm OS. On the first postoperative day, the patient's uncorrected visual acuity was 20/40 OD and 20/60 OS, but at the next examination 6 weeks later, it had decreased to 20/400 OU, and his corrected visual acuity was 20/400 OD and 20/30 OS with a refraction of –5.50 sphere OU. A bilateral simultaneous enhancement procedure was performed under the flaps to fully correct the estimated residual myopia, and on the next examination 2 weeks later, the patient's uncorrected visual acuity was 20/400 OD and 20/40 OS. Six weeks later, his best spectacle-corrected visual acuity was 20/25 OU with a refraction of –1.50 – 3.00 × 77 OD and –2.50 –0.75 × 170 OS.

During the next 10 months, the refractive error of the patient's right eye progressed, necessitating frequent spectacle changes, from –3.25 –4.00 × 66 to –11.00 – 4.75 × 71. His best spectacle-corrected visual acuity decreased to 20/200 OD. Subjectively, the left eye remained relatively stable. On examination approximately 1 year following the initial procedures, his best spectacle-corrected visual acuity was 20/100 OD and 20/30 OS with a refraction of –10.50 − 4.00 × 67 OD and –3.75 sphere OS. Findings from slitlamp examination revealed a hinged keratomileusis flap in both corneas and ectasia of the right cornea with a steep central protrusion. Ultrasound pachymetry measurements were 386 µm OD and 485 µm OS. A computed corneal topographic map of the right eye showed profound inferonasal steepening in a keratoconus pattern with an apical corneal power in excess of 50 D (Figure 2), while a map of the left eye showed a well-centered excimer laser ablation with marked flattening. The patient's right eye was fitted with a rigid contact lens, with which he achieved satisfactory visual acuity.

Keratoconus is generally considered a contraindication for excimer laser refractive surgery, since it is expected that the progression of ectasia is likely to be hastened by the removal of central corneal tissue. Some authors have suggested that the risk is greatly exaggerated and have reported no evident acceleration of ectasia from 6 to 46 months after performing surface photorefractive keratectomy (PRK) on patients with a clinical diagnosis of keratoconus.¹ While experience is limited and follow-up brief, these results have prompted others to perform PRK in eyes that might be classified as forme fruste keratoconus; ie, those eyes demonstrating topographic changes suggestive of keratoconus but without notable thinning, ectasia, or scarring. It might be argued that such corneas might share biomechanical properties with those that demonstrate true keratoconus, and so demonstrate an increased tendency toward progression of thinning, myopia, and astigmatism after surgery. Indeed, Kremer et al² have demonstrated in a study of 8 eyes in 6 patients with mild keratoconus (compound myopic astigmatism and topographic features consistent with keratoconus without notable ectasia, thinning, or scarring) who were followed for more than 3 years after undergoing surface PRK that while most experienced improvement in unaided vision, 1 suffered progression of keratoconus. Conversely, Doyle et al,³ arguing that topographic evidence of inferior corneal steepening in the absence of clinical signs consistent with keratoconus is often artifactitious, performed PRK in 4 such eyes and found the results comparable to those expected in normal eyes with myopic astigmatism.

Therefore, the risk of a poor visual outcome and progressive ectasia after performing surface PRK in eyes with isolated topographic abnormalities suggestive of keratoconus is unclear. Although there is evidence to suggest that performing surface PRK might be successful in select cases, it cannot be assumed that this experience is applicable to patients undergoing LASIK procedures in which ablation is performed in a partial-thickness corneal bed, producing a relatively thinner effective stress-bearing cornea. Our patient provides an instructive case-control study of the risk of ectasia following LASIK in eyes with no observable signs of keratoconus but in which corneal topographic scan findings are suspect. While neither eye demonstrated frank keratoconus, progressive ectasia occurred exclusively in the eye with a suspect corneal topography. Previous reports are few. Seiler and Quurke⁴ have also described progressive corneal ectasia that occurred in a patient with an asymmetric bow-tie pattern that they interpreted as forme fruste keratoconus. In our case, in addition to an asymmetric bow-tie pattern, other features suggested subclinical keratoconus, including markedly asymmetric topographic scan findings between the 2 eyes and poor initial best spectacle-corrected visual acuity.

It might be postulated that the risk of ectasia following LASIK might be higher than that following PRK because of the relatively thinner effective stress-bearing corneal stroma, but the residual stromal-bed thickness required to avoid progressive corneal ectasia in either topographically normal or abnormal eyes that undergo LASIK is unknown. Based on personal experience, Barraquer⁵ has suggested a 300-µm thickness of stress-bearing cornea. By comparing the biomechanical properties of keratoconic corneas with normal corneas, Andreassen et al⁶ have estimated that for the normal cornea, a residual stromal-bed thickness of less than 250 µm might produce a cornea with a tangential elastic modulus comparable to that of a keratoconic cornea.

Lyle and Jin⁷ have reported a high incidence (26%) of progressive corneal ectasia that they termed iatrogenic keratoconus following hyperopic automated lamellar keratoplasty. The depth of the lamellar cut in this patient group ranged from 52% to 70%. However, their cases included corneas that had undergone prior radial keratotomy and were therefore structurally weakened to begin with. Hyperopic lamellar surgery frequently produces residual stromal beds less than 250-µm thick, but when performed in cases of primary hyperopia rather than consecutive hyperopia following radial keratotomy, progressive ectasia is relatively rare.⁸ In 3 cases of progressive corneal ectasia and keratoconus-like steepening developing in topographically normal-appearing eyes that underwent LASIK, Seiler et al⁹ estimated that the residual stromal-bed thickness was less than 200 µm in 1 patient and between 200 µm and 250 µm in the others. Based on this experience and on the theoretical calculations of Andreassen et al,⁶ these authors advocated a minimal residual stromal-bed thickness of 250 µm.

In the case we report, the estimated residual stromal-bed thickness prior to enhancement was 260 µm OD. This suggests that for certain corneas, such as those demonstrating features of keratoconus on topography, even 250 µm may not be an adequate stromal-bed thickness to prevent progressive ectasia. Alternatively, the actual thickness of the flap created by the microkeratome might differ substantially from the expected, which might produce a thinner than anticipated bed.¹⁰ Also, reading error in pachymetry measurements obtained, for example, from thicker paracentral regions of the cornea might also have led to overestimation of the residual stromal-bed thickness.

While these cases of corneal ectasia following LASIK have demonstrated fairly rapid progression, we are concerned that others may develop more slowly. In a study of loss of refractive effect in the first year following LASIK, Chayet et al¹¹ report that regression seemed to be caused by an increase in corneal thickness rather than ectasia. However, longer-term studies are necessary. We strongly advocate that until we are better able to identify patients at risk for ectasia following LASIK, and the variables defining the biomechanical properties of the operated cornea are better described, LASIK should not be performed when findings from the examination or corneal topography suggest subclinical keratoconus.

Corresponding author: Stephen D. McLeod, MD, Department of Ophthalmology, University of California–San Francisco, 10 Kirkham St, K-301, San Francisco, CA 94143, (e-mail: smcleod@itsa.ucsf.edu).