A review of complications associated with laser in situ keratomileusis
(LASIK) indicates that most are directly attributable to the creation of a
corneal flap.1,2 In their
examination of 1000 consecutive cases of patients who had undergone LASIK,
Gimbel and colleagues1 identified 32 intraoperative
and 18 postoperative complications, most of which could be related to issues
of flap anatomy, including incomplete passes, thin flaps, buttonholes, flap
shrinkage and flap dislocation with subsequent development of striae, and
epithelial ingrowth. Stulting and colleagues2
reported complications encountered in a series of 1062 cases of patients who
had undergone LASIK, and identified 27 intraoperative and 40 postoperative
complications, all of which were directly related to the corneal flap with
the exception of 2 cases of keratitis.
Although most complications can be resolved with acceptable visual outcomes,
persistent flap irregularity or opacification will result in decreased vision.
Since epithelialization of the underlying stromal bed might provide a more
regular surface, amputation of the offending flap might be considered a reasonable
intervention to address persistent flap problems. It is therefore important
to understand the healing pattern of the corneal bed following flap creation
and excimer laser ablation in terms of lens power, topography, regularity,
and scar formation. In our experience, most cases of flap amputation have
followed infectious keratitis and flap melting that results in some degree
of scarring and opacification of the underlying corneal bed. Consequently,
it has been difficult to predict what the optical qualities of the uninflamed
stromal bed might have been. We document herein the corneal findings in 2
patients who underwent early flap amputation for noninflammatory epithelial
ingrowth following LASIK.
A 46-year-old woman with a history of recurrent corneal erosion and
an examination finding consistent with map-dot-fingerprint dystrophy underwent
bilateral LASIK for the correction of an error of –6.75 + 0.50 ×
100 OD and –7.00 + 0.25 × 072 OS. An automated microkeratome (Automated
Corneal Shaper [ACS]; Bausch & Lomb Surgical, Rochester, NY) was used
to create the corneal flaps with nasally located hinges, followed by ablation
with an excimer laser (VISX Star; VISX, Inc, Santa Clara, Calif). An epithelial
defect was produced during surgery in the left eye. A bandage soft contact
lens was placed, but a defect persisted at the first follow-up visit 1 day
later. Approximately 3 weeks later, epithelial ingrowth along the interface
of the corneal flap and the bed was identified at the hinge and extended toward
the entrance pupil. The flap was elevated and the interface epithelium, removed.
Approximately 2 weeks later, the epithelial ingrowth had recurred, so the
flap was amputated.
The patient was treated with a bandage contact lens, and ciprofloxacin
hydrochloride solution was applied every 3 hours. No corticosteroids were
applied. During the next 5 days, the epithelial defect created by removal
of the flap closed, the bandage contact lens was removed, and the patient
was prescribed diclofenac sodium solution for occasional use up to 3 times
daily and artificial tears for lubrication. Approximately 1 week after closure
of the epithelial defect, the uncorrected visual acuity in the left eye was
20/100. Automated refraction identified an error of –5.50 + 3.50 ×
159, but the corresponding visual acuity was not recorded. Topical corticosteroids
were prescribed for application 3 times daily and discontinued after 1 month.
During the next 6 months, the corneal haze in the left eye was not recorded
as being any greater than 1+. However, at 9 months after flap amputation,
the uncorrected vision was recorded as 20/100, correcting to 20/40 with a
refraction of –1.50 + 0.25 × 171.
The patient was referred to the University of California, San Francisco
Refractive Surgery Service for further consultation in May 2001, approximately
2 years after LASIK and flap amputation of the left eye. At that time, she
complained of fluctuating vision in the left eye that at its best remained
blurred. She also reported ghosting and glare. Examination disclosed an uncorrected
visual acuity of 20/25 OD and 20/80 OS. The vision of the left eye improved
to 20/25 with a refraction of –3.25 + 3.25 × 70.
Slitlamp biomicroscopic examination of the right eye showed a well-positioned,
nasally hinged corneal flap, but coarse, diffuse epitheliopathy. There was
no evidence of subepithelial or stromal haze or scarring, except for a normal
degree of scar formation outlining the edges of the corneal flap. Slitlamp
examination of the left eye showed a subtle, vertically oriented elevation
of the corneal surface at the hinge of the amputated flap. There was no evidence
of subepithelial or stromal scarring, either at the former location of the
flap edge or over the central cornea. However, there was a moderate degree
of epithelial irregularity evident without instillation of fluorescein sodium
dye. Fluorescein sodium staining revealed coarse, diffuse epitheliopathy concentrated
over the central cornea and an area of irregular surface contour that appeared
to involve the central area of corneal dissection that produced the amputated
Computerized corneal mapping (Figure
1) confirmed the relative irregularity of the left eye. A topographic
map of the right eye (Tomey Topographic Modeling System, version 2.3.6J; Tomey
Corp, Waltham, Mass) showed a simulated keratometry reading of 40.01 ×
41.07@91° with a surface regularity index of 0.52 and a surface asymmetry
index of 0.18. However, a topographic map of the left eye produced a simulated
keratometry reading of 40.90 × 42.51@103° with a surface regularity
index of 0.50 and a surface asymmetry index of 1.25. On comparing the right
and left eyes, a markedly asymmetrical reflex was also observed on retinoscopy
with significantly greater irregularity noted in the left eye.
Case 1. Corneal topography of
both eyes. Although the overall powers of the central corneal curvatures are
similar, the left eye shows greater irregularity, as represented by the elevated
surface asymmetry index (SAI). SimK indicates simulated keratometry; MinK,
minimum keratometry; PVA, predicted visual acuity; CYL, cylinder; and SRI,
surface regularity index.
Although a relatively high degree of astigmatism was noted in the left
eye, the spherical equivalent was calculated to be –1.625 diopters (D).
Since the refraction in the right eye was –1.50 + 1.00 × 090,
anisometropia was limited, so spectacles were prescribed to improve visual
A 33-year-old man underwent bilateral LASIK for the correction of an
error of –1.75 + 0.50 × 30 OD and –2.00 + 0.25 × 160
OS. An automated microkeratome (ACS; Bausch & Lomb Surgical) was used
to create the corneal flaps. A large epithelial defect was created in the
left eye, so the flap was repositioned without excimer laser ablation. A bandage
contact lens was placed to promote epithelial healing. Approximately 2 months
later, the patient returned to surgery. A corneal flap with a nasal hinge
was created in the left cornea using an automated microkeratome (ACS; Bausch
& Lomb Surgical), and the ablation was performed using an excimer laser.
An epithelial defect was noted at the end of the procedure, and a bandage
contact lens was kept in place for the next 3 days. One week after surgery,
uncorrected vision was 20/40 OS, correcting to 20/25 with a refraction of
–1.00 + 1.50 × 20. No significant epithelial ingrowth was noted.
Two weeks later, the patient returned with the complaint of ocular discomfort
in the left eye. Uncorrected vision was 20/30−. Epithelial ingrowth
was noted along the nasal hinge, with extension toward the entrance pupil.
At that visit, the flap was lifted to remove the interface epithelium, and
the epithelium overlying the flap was noted to be friable. On the basis of
anticipated difficulties with recurrent epithelial ingrowth, the flap was
amputated and a bandage contact lens was placed. Ciprofloxacin and diclofenac
solutions were prescribed 4 times daily. The epithelial defect healed during
the next few days, and 1 week after flap amputation the uncorrected visual
acuity was 20/200, correcting to 20/60 with a refraction of –5.00 +
1.50 × 100. The ciprofloxacin solution was discontinued, and corticosteroid
drops were prescribed for use 3 times daily. Two months later, the uncorrected
vision was 20/100, correcting to 20/50− with a refraction of –4.75
+ 2.00 × 105. The corticosteroid therapy was reduced to 1 drop per day
and then discontinued.
The patient complained of poor vision and nighttime glare and halo and
was referred to the University of California, San Francisco Refractive Surgery
Service for consultation in May 2001, approximately 18 months after LASIK
and subsequent flap amputation of the left eye. Examination at that time disclosed
an uncorrected visual acuity of 20/25 OD, correcting to 20/20 with a refraction
of –0.50 + 0.50 × 55, and an uncorrected visual acuity of 20/60
OS, correcting to 20/20− with a refraction of –3.75 + 3.75 ×
97. Pachymetry readings were 532 µm OD and 439 µm OS.
Slitlamp biomicroscopic examination of the right eye showed a well-positioned,
nasally hinged corneal flap with mild central subepithelial opacification,
whereas slitlamp examination of the left eye was remarkable for mild vertical
linear elevation at the site of the transected hinge, a semicircle of subepithelial
haze reminiscent of surface photorefractive keratectomy (PRK)–associated
scarring that appeared to outline the perimeter of the flap, and a relatively
lucent central cornea overlying the entrance pupil. (Figure 2) The surface of the central cornea appeared to be relatively
smooth, but upon instillation of fluorescein sodium solution, inspection of
the tear film pattern indicated an irregular surface.
Case 2. Slitlamp photograph of
the left eye. Note within the slit beam a band of scarring that outlines the
perimeter of the flap, but relative clearing of the central cornea overlying
the entrance pupil.
The irregularity of the left eye's corneal surface was confirmed by
computerized corneal mapping (Figure 3).
A topographic map of the right eye produced a simulated keratometry reading
of 42.27 × 43.32@84°, with a surface regularity index of 0.11 and
a surface asymmetry index of 0.50. However, a topographic map of the left
eye showed a simulated keratometry reading of 43.57 × 46.40@115°,
with a surface regularity index of 1.62 and a surface asymmetry index of 0.86.
On comparing the right and left eyes, a markedly asymmetrical reflex was also
observed on retinoscopy, with markedly greater irregularity noted in the left
Case 2. Corneal topography of
both eyes. The left eye shows significant astigmatism with the rule and elevated
irregularity compared with the right eye, as represented by elevated surface
regularity (SRI) and surface asymmetry indices (SAI). Other abbreviations
are explained in the legend to Figure 1.
Since the acuity in the left eye could be corrected to 20/20−
with a relatively low degree of anisometropia based on spherical equivalent,
spectacles were recommended, but the patient adamantly refused to consider
spectacle correction. Rigid contact lenses were also suggested, but the patient
elected to forgo fitting.
The findings from large reported series of complications seen in consecutive
cases of patients who have undergone LASIK suggest that most complications
can be attributed to abnormalities of the corneal flap that translate to irregularity
or opacification of the anterior cornea.1,2
If amputation of the corneal flap were followed by reepithelialization and
smoothing of the corneal surface (analogous to corneal healing after surface
PRK) without the introduction of significant scarring, refractive error, or
irregularity, then this approach might prove useful in addressing most postsurgical
complications of LASIK. Unfortunately, few reports in the literature provide
a guide to the clinical course that can be expected after flap amputation
in the uninfected cornea. Patel and colleagues3
recently reported a case of traumatic flap dislocation that was followed by
loss of the flap. After healing of the stromal bed, the patient's uncorrected
vision was 20/40, improving to 20/20 with a refraction of –1.00 + 1.00
However, as our 2 cases demonstrate, it cannot be assumed that a regular
surface will result after removal of the flap. The irregular myopic astigmatism
we observed implies that the curvature of the stromal bed might not precisely
reflect that of the anterior surface of the overlying flap. This finding suggests
that the flap might vary in thickness from one region to another, leading
to variability in the curvature of the stromal bed created. Patterns of variability
in thickness may well differ from one microkeratome to another, and this variability
is expected to contribute to the development of irregularity in the contour
of the stromal bed, which results in irregular astigmatism that limits best
Under normal circumstances, irregularity of the surface of the stromal
bed is expected to be matched by corresponding irregularity of the undersurface
of the flap, so that if an irregular flap is created and then replaced precisely
with a "lock and key" effect, little change on the anterior corneal surface
is expected. If a regular refractive ablation is performed on the exposed
stromal bed, some degree of underlying irregularity should be translated through
the ablation so that as the surfaces are reapposed precisely, matching the
bed to the underside of the flap, the composite effect on the surface of the
eye should be attenuation of the irregularity. As our 2 cases suggest, this
attenuating effect is lost if the flap is removed to expose the stromal bed.
Over time, remodeling of the epithelium might have a smoothing effect
on the exposed stromal bed, improving best spectacle-corrected visual acuity.
After this improvement, residual regular spherocylindrical error can be corrected
with spectacles, hydrophilic contact lenses, or a standard spherocylindrical
excimer laser treatment. In a topographic examination of eyes treated with
excimer laser, Abbas and colleagues4 have
demonstrated that significant corneal smoothing occurs from 3 months to 12
months after PRK, presumably as a result of stromal healing and remodeling.
Using very high-frequency ultrasound scanning, Reinstein and colleagues5 examined corneas that had undergone LASIK and reported
regional variations in epithelial thickness that tended to compensate for
underlying stromal irregularity, thereby reducing corneal irregularity. In
the 2 cases we present, it is discouraging that reduced best spectacle-corrected
visual acuity with correspondingly elevated indices of asymmetry and irregularity
was evident 18 months and 2 years after flap amputation. Therefore, it is
questionable how much further improvement in surface regularity might occur
during subsequent months or years.
In neither case was there substantial scarring of the corneal stroma
overlying the entrance pupil that was subjected to excimer ablation. High
degrees of refractive error corrected by surface PRK are expected to be associated
with a greater risk of scarring, and it has been suggested that this scarring
is related to the depth of the ablation performed.6
However, after flap amputation, relatively deep layers of the cornea were
exposed to the epithelium after healing, and no significant haze was recorded
throughout the healing period. This finding suggests that the risk for haze
formation in PRK probably goes beyond simple considerations of exposure of
the deeper stroma devoid of Bowman membrane to healing epithelium. Rather,
these cases suggest that flap amputation is not necessarily followed by significant
central corneal haze and scarring.
Nevertheless, the refractive and topographic outcomes of our 2 patients
indicate that there is a substantial risk for refractive change and induction
of irregular astigmatism following flap amputation. Any characteristic pattern
of induced astigmatism is probably related to the path followed by the microkeratome
in creating the flap, which in turn will be related to the particular design
of the microkeratome. Since there were no other flap-related abnormalities
beyond epithelial ingrowth in these cases, we surmise that flap amputation
was performed because it was seen as a definitive treatment of the ingrowth
that would produce acceptable surface smoothing over time.
The first patient we describe had a history of recurrent erosion syndrome,
which presents an increased risk for epithelial ingrowth, keratolysis, flap
melting, and loss of best corrected visual acuity.7
For this reason, LASIK is not recommended in the setting of anterior basement
membrane disease, and PRK should be considered. Such severe complications
might indeed ultimately necessitate flap amputation, but no such progression
was seen in the cases reported herein. Therefore, based on the observed long-term
clinical course, we suggest that in the absence of compelling indications
(such as gross flap irregularities or interface infection in which the flap
might limit antibiotics penetration), flap amputation should be a last resort
in the management of flap complications.
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: firstname.lastname@example.org).
McLeod SD, Holsclaw D, Lee S. Refractive, Topographic, and Visual Effects of Flap Amputation Following Laser In Situ Keratomileusis. Arch Ophthalmol. 2002;120(9):1213-1217. doi: