Corneal haze is a vision-threatening complication of photorefractive keratectomy (PRK). Haze is the result of an inflammatory healing response to PRK-induced corneal trauma that involves the migration and replication of keratocytes and myofibroblasts.1,2 Mitomycin-C (MMC) is a potent antineoplastic agent that inhibits proliferation of all cell types, including keratocytes and myofibroblasts, in the bathed cornea. Perioperative use of MMC has been shown to reduce the risk for corneal haze; however, few prospective studies have evaluated its use with modern ablation profiles.3–6
The purpose of this study was to assess the safety and efficacy of MMC 0.01% for the prevention of haze after wavefront-guided PRK at 3 dose durations: 60 seconds, 30 seconds, and 15 seconds. The objectives included to determine (1) whether MMC at 0.01% will prevent corneal haze after PRK for high myopia at the 3 decreasing durations, (2) the differences in the incidence of visually significant corneal haze between the treated eye and the untreated control eye, (3) the differences in refractive outcomes and achievement of intended refraction between the treated eye and the control eye, and (4) differences in endothelial cell density (ECD) between the treated eye and the control eye.
Patients and methods
The study was approved by the hospital’s Institutional Review Board (CIP #NMCSD.2006.0048), and an Investigational New Drug (IND) request for off-label use of MMC was filed with the U.S. Food and Drug Administration (FDA) (FDA IND 76 601). Patients were recruited from those seeking PRK at the Naval Medical Center. Informed consent was obtained from all participants.
Patients electing to have PRK who were at least 21 years old with a manifest spherical equivalent (SE) from −4.50 diopters (D) to −9.00 D with refractive cylinder of up to 3.00 D were offered enrollment. Patients were required to have a corrected distance visual acuity (CDVA) with spectacles of 20/25 or better in both eyes and a stable spectacle refraction, confirmed by clinical records documenting that neither the spherical nor the cylindrical portion of the refraction had changed by more than 0.50 D during the 6-month period immediately preceding the baseline examination. No more than 1.00 D of anisometropia of the SE was allowed so that ablation depths between the 2 eyes would be relatively equivalent. Naval aviators, patients who were pregnant or breastfeeding, patients with preexisting corneal opacities, and patients with any FDA medical contraindication to corneal refractive surgery were excluded.
Treatment Group Assignment and Randomization
There were 3 treatment groups in this study: Group A: 60 seconds of MMC exposure; Group B: 30 seconds of MMC exposure; Group C: 15 seconds of MMC exposure. Patients were distributed alternately to treatment Group A, B, or C as they enrolled. In addition, patients were randomized as to which eye received MMC and which received a placebo according to a randomization list prepared in advance for each treatment group. The randomization list was generated by Excel software (Microsoft Corp.) using the =Rand() formula. The surgical team was aware of which eye received the MMC; however, the patient and the optometrists providing the follow-up visits were masked to which eye received the medication. After surgery, the operative note outlining which eye received which treatment was placed in a sealed envelope in the patient’s chart. The randomization could be revealed at study completion, when the patient developed an adverse event, or when the patient developed haze severe enough that surgical treatment was being considered, whichever came first. Medical treatment of haze with topical steroid drops was not a criterion for revealing the randomization.
All patients had a standard refractive surgery preoperative evaluation that included manifest refraction, CDVA (4 m Early Treatment of Diabetic Retinopathy Study [ETDRS] self-illuminated eye chart, Lighthouse 2nd ed.), cycloplegic refraction, slitlamp examination, dilated fundus examination, corneal ultrasound pachymetry, wavefront aberrometry (Wavescan, Visx, Inc.), and corneal topography (Humphrey Atlas, Carl Zeiss Meditec AG). In addition to the standard preoperative examination elements, corrected photopic contrast sensitivity was measured using a 5% contrast back-illuminated contrast sensitivity chart (Precision Vision), and mesopic (approximately 1 candela/m2) contrast sensitivity was measured in a dark room using a back-illuminated 25% contrast chart covered with a 2-log neutral density filter supplied by the manufacturer (Precision Vision). Central cornea endothelial cell count (ECC) images were obtained during 2 separate preoperative visits using the Nidek confocal microscope (Confoscan 4, ×20 magnification) and nontouch automated counting of approximately 400 to 650 endothelial cells.
Bilateral wavefront-guided PRK was performed using a standard technique and the VISX Star S4 laser (Customvue, IR, Abbott Medical Optics, Inc.). After anesthesia of topical proparacaine was applied, a lid speculum was placed and an Amoils epithelial scrubber (Innova, Inc.) was used to remove the corneal epithelium before laser treatment. Immediately after the excimer laser treatment, the cornea was irrigated with a balanced salt solution. Then, the ocular surface was carefully dried with a polyvinyl alcohol sponge (Merocel). A precut 6.0 mm diameter polyvinyl alcohol sponge was then saturated with MMC or a balanced salt solution according to randomization. The following technique was used to saturate the MMC sponge: A 0.3 mL insulin syringe was used to apply 0.04 mL of MMC 0.01%, (0.1 mg/mL) solution. This volume delivered 0.004 mg of MMC and saturated the sponge to the point that it was wet but not dripping. The moistened sponge was not squeezed after study medication had been applied to it. The moistened sponge was then applied to the cornea for 15 seconds, 30 seconds, or 60 seconds depending on the treatment group. The ocular surface was rinsed with the entire contents of a 15 mL bottle of a balanced salt solution. The sponge was then disposed of in hazardous waste bags according to standard protocol. A bandage contact lens (Ciba Focus Night and Day) was placed, and topical ketorolac and gatifloxacin were administered.
All patients received standard PRK postoperative medications that included topical gatifloxacin while the bandage contact lens was in place, topical lubricant eyedrops, oral ibuprofen as needed, and oral acetaminophen–oxycodone as needed. A 16-week tapering dose of fluorometholone was prescribed in sequence as follows: 4 times a day for 4 weeks, 3 times a day for 4 weeks, 2 times a day for 4 weeks, and 1 time a day for 4 weeks. Any eye that developed visually significant haze could be treated according to the center’s standard protocol, which was to increase the topical fluorometholone or to switch to topical prednisolone depending on the severity of the haze. Patients were also given ultraviolet (UV) protective sunglasses in their postoperative kits and were instructed to wear them whenever they were outside during daylight; however, no data on compliance with UV protection was collected.
Patients were evaluated on the first postoperative day, at 4 to 7 days, at 2 weeks, and at 1, 3, 6, and 12 months. The bandage contact lenses were removed 4 to 7 days postoperatively. Refractions, acuities using ETDRS letter charts (100% and 5% photopic and 25% mesopic), and a slitlamp examination with assessment of corneal haze were obtained at every visit. Haze was scored according to the corneal haze grades established by Braunstein et al.7 (Table 1). Endothelial cell counts were obtained at the 1-, 6-, and 12-month postoperative visits using the same confocal microscope as preoperatively.
The main efficacy variable in the study was the presence or absence of corneal haze. The analysis consisted of a 2 × 3 × 3, 4, or 6 repeated measures (depending on the variable) analysis of variance (ANOVA) with interaction and with post hoc multiple comparisons. The dimensions represented 2 treatment categories (with MMC and without MMC), 1 for each eye (nonrepeated); 3 durations of MMC exposure (nonrepeated); and 3 (or more) postoperative times of haze rating, ECD, or visual performance measurement (repeated). A power analysis based on patient charts from agglomerated records using an α equal to 0.05, power equal to 0.90, and standard deviations of haze ratings of 0.84 and 0.37 for the untreated and treated patients, respectively, indicates that a haze rating difference of 1 unit could be detected with 9 patients in each of the 3 dosage groups (18 eyes). All statistical analyses were performed using Statistica software (version 9.1, Statsoft).
Twenty-eight patients were enrolled: 10 in the 60-second exposure to MMC group, 9 in the 30-second exposure group, and 9 in the 15-second exposure group. Table 2 shows the patients’ baseline demographics. There were no statistically significant differences in the age, treatment wavefront SE, or ablation depth between the treatment groups (ANOVA; for age, F = 0.008; for SE, F = 0.517; for depth, F = 2.039; all P > .05). There was a difference in preoperative cylinder between Group B (30 seconds MMC) and Group C (15 seconds MMC) (−1.90 D versus −0.74 D; P=.047, Tukey honesty significant difference [HSD]), but not between Group A (60 seconds MMC) and Group C (P=.080, Tukey HSD).
Biomicroscopic examination showed that no eye developed more than trace (1.0) haze at any postoperative visit. No eye had visual performance–altering haze. There was no difference in any visual acuity outcomes between the MMC eye and the untreated control eye in any group at any postoperative visit. One MMC eye (4%) had trace haze (score = 1), which was seen during the 6-month postoperative examination. Three contralateral control eyes (11%) had trace haze (score = 1) at the 1-month examination, 1 (4%) at the 3-month examination, and 1 (4%) at the 6-month examination. One patient had the bandage contact lenses replaced on the first postoperative day and was given an extra dose of steroids as part of that procedure. Other than this instance, no eye required additional steroids. There was a significant, but slight difference in haze scores between MMC eyes and control eyes at the 1-month visit and 3-month visit (P=.028, Wilcoxon matched-pair test); however, there was no difference at the 6-month and 12-month examinations (Figure 1). By 12 months, all haze scores were zero.
Endothelial cell densities decreased slightly in the MMC eyes and control eyes in all 3 exposure groups 1 month postoperatively (Figure 2). The density of cells was not influenced by the duration of exposure to MMC (F = 0.197, P=.82) or whether the eye was exposed to MMC (F = 0.223, P=.64). Six and 12 months after treatment, the ECD in the central region of the cornea was the same as the preoperative ECD (P > .30, t test). One month after treatment, the ECD was reduced by a mean of 99.6 ± 192 cells/mm2 over a mean preoperative density of 2496 ± 242 cells/mm2 (P=.0004, t test).
There were no significant differences in the postoperative visual performance as the result of MMC exposure. In a multivariate ANOVA, postoperative visual performance was not predicted by whether the eye was treated with MMC or the duration of MMC exposure. Visual performance was represented by the efficacy (postoperative uncorrected distance visual acuity [UDVA] relative to preoperative CDVA) (Figure 3), manifest refraction (Figure 4), corrected photopic acuity (Figure 5), or corrected mesopic contrast acuity (Figure 6). The individual F statistic for MMC exposure or treatment duration, or the interaction between exposure and duration, was never greater than 1.84; thus, all probabilities were greater than 0.18.
For efficacy, no difference was found as a function of duration of MMC exposure (F = 0.996, P=.326) or whether the eye was treated with MMC or a placebo (F = 0.053, P=.819). Also, the postoperative manifest refraction was not significantly influenced by the duration of MMC exposure (F = 0.016, P=.90) or whether eyes were exposed to MMC (F = 1.24, P=.30). The CDVA was preserved postoperatively but was not significantly influenced by duration of exposure to MMC (F = 1.84, P=.176) or whether MMC was used during treatment (F = 0.292, P=.592). Mesopic contrast acuity was also well preserved and was not influenced by the duration of MMC exposure (F = 1.74, P=.19) or whether MMC was used (F = 0.23, P=.64).
One patient developed an epithelial erosion 10 weeks (73 days) after surgery; the erosion was unrelated to trauma. The treatment randomization for that patient was unmasked and it was determined that MMC had not been applied to the eye with the erosion. The erosion healed quickly with additional topical lubricants. This eye did not develop haze, and there was no recurrence of the abrasion. No extra steroids were given to aid with recovery.
Haze after PRK, although rare, can cause significant loss of CDVA. The epithelial defect, photoablation of the stromal surface, and possible thermal damage from the excimer ablation may all play a role.8–12 Several studies note that the deeper ablation associated with high levels of myopia, hyperopia, or astigmatism are risk factors for haze formation.13 Age does not appear to be associated with haze, nor does keloid formation; however, there may be racial factors.14–16
Other proposed risk factors include more intense exposure to the sun’s UV light and noncompliance with postoperative steroid regimen.17 Carones et al.3 published the first prospective evaluation of prophylactic use of MMC. They found that MMC 0.02% for 2 minutes essentially eliminated the risk for post-PRK haze in high myopia with conventional excimer laser treatment profiles. Before our current study, our standard practice was to apply MMC in half the concentration described in the literature (ie, 0.01%) for 1 minute after PRK in patients who were considered at high risk for haze development. High risk was defined as patients requiring deeper ablations (>50 μm) and patients whose jobs or recreational activities caused them to have significant exposure to ambient UV light. Anecdotally, after this MMC treatment regimen was implemented at our center, the incidence of haze after PRK was greatly reduced. However, during the same time, we also switched from conventional excimer laser treatment profiles on the VISX S4 excimer laser to wavefront-guided procedures on the VISX S4 excimer laser. A wavefront-guided procedure purportedly has a smoother ablation profile and a more gently tapered transition zone. Therefore, it was not clear whether the reduction in haze in our practice was secondary to MMC use or the wavefront-guided ablation profile. We believe that this is the first prospective placebo-controlled trial evaluating a dose-response for MMC 0.01% in wavefront-guided PRK for high myopia.
Since Carones et al.’s3 landmark article, several studies have evaluated MMC 0.02% with PRK. Thornton et al.18 report a reduction of haze with the inadvertent application of even lower concentrations of MMC (ie, 0.002%). Thornton et al.’s retrospective study found that using this very low dose of MMC for 30 or 120 seconds resulted in 5% to 10% of eyes having a haze score of more than 1 throughout a 1-year follow-up. A case-control group of patients with similar refractive errors who received that center’s standard dose of MMC (0.02%, 30 or 120 seconds) had no haze score of more than 1. Leccisotti19 found similar results in a study of MMC and hyperopic PRK, with 11% of placebo eyes having a haze score of more than 1 and MMC-treated eyes (0.02% for 45 seconds) having scores of 1 or less. Virasch et al.20 examined this question retrospectively and found that a 12-second application of MMC 0.02% was just as effective as longer durations (60 seconds and 120 seconds).
In addition to delineating the proper dose of MMC for PRK haze prevention, showing the safety of MMC in these healthy eyes was a paramount concern. Mitomycin-C has been in use in ophthalmologic surgery since the late 1980s. In addition to prevention of haze after PRK, its various indications include prevention of scarring after trabeculectomy (glaucoma filtering surgery), prevention of recurrence of pterygia, and treatment of conjunctival and corneal squamous cell dysplasia and carcinoma. However, in high doses, MMC can cause devastating complications in the eye. The complications can appear immediately or insidiously over the first postoperative weeks. One serious complication is nonhealing conjunctival and scleral ulceration, which can lead to loss of the eye. Complications during the initial use of MMC in ophthalmologic surgery were seen in patients treated with higher doses than are in use today, such as 0.04 %, or with prolonged postoperative topical use. There have also been concerns that MMC may be toxic to the corneal endothelium. Complete endothelial cell loss requiring corneal transplantation was seen in a patient who was treated with topical MMC drops after phototherapeutic keratectomy, a surface excimer laser procedure similar to PRK that does not include a refractive component to the ablation.21 We found it reassuring that MMC did not adversely affect the ECD in any eye in our study. Both control eyes and MMC-treated eyes showed a temporary reduction in ECD, which recovered over time.
A review of the literature showed 7 applicable studies with contradictory conclusions regarding the effect of MMC on the endothelium.5,19,22–26 Goldsberry et al.23 report no loss of ECD in a small group of 16 eyes with the use of MMC 0.02% for PRK for high myopia using a short exposure (12 seconds). There was no influence on ECD22,23 or very limited loss of cell density21 with the application of MMC 0.02% for 15 seconds. When MMC 0.20% was applied for 30 seconds or longer, some studies21,25 found a decrease in ECD. Nassiri et al.25 report a loss of ECD of approximately 400 cells/mm2 ± 166 (SD) with the application of MMC 0.02% for 10 to 50 seconds. Untreated control eyes with a similar refractive error lost a mean of approximately 120 ± 122 cells/mm2. The study was limited because ECCs were measured through 24 weeks postoperatively only and the exposure of MMC was calculated, in steps, as a function of the intended ablation depth. Figure 7 compares our study results with those available in the current literature addressing the influence of MMC on ECD at different concentrations and durations of exposure.5,19,22–26
There are several limitations to this study. A 4-month postoperative steroid regimen could have been a significant factor in the small difference in haze between the MMC eye and the control eye. Similar results were found in a study by Gambato et al.27 that compared MMC 0.02% for 2 minutes with a 3-month taper of fluorometholone ophthalmic suspension 0.1%. In early PRK studies, steroids appeared to reduce haze formation with smaller ablation zones.28,29 It is not clear what role topical steroids play in the amelioration of haze development with modern ablation profiles. This was not analyzed in our study. In addition, although the San Diego area enjoys many months of sunshine per year, it is not known whether our results would be applicable to patients treated at other centers who may be exposed to more intense, naturally occurring UV light.
In conclusion, in this group of patients with higher myopia who had PRK in San Diego with a 4-month postoperative steroid regimen, there was no clinically significant difference in haze formation between eyes that received MMC at any dose and the fellow control eye. Mitomycin-C had no adverse effect on the ECC, with both MMC-treated eyes and control eyes having a slight decrease 1 month postoperatively; this recovered by 6 months. The use of MMC appears to be safe; however, the previously held ideas about high myopia being a risk factor for haze may no longer be true with modern excimer ablation profiles. A larger study involving a shorter course of steroids and lower refractive errors may help answer this complex question.
What Was Known
- Use of MMC 0.02% (0.2 mg/mL) applied for 2 minutes after conventional PRK for high myopia reduces the incidence of post-PRK haze.
What This Paper Adds
- A lower concentration of MMC 0.01%, (0.1 mg/mL) applied for shorter durations than previously described appeared to be effective at preventing post-PRK haze in patients with high myopia having wavefront-guided PRK.
- The wavefront-guided ablation itself may play a role in decreasing post-PRK haze, as evidenced in a very low incidence of post-PRK haze in control eyes.
1. Kim T-I, Pak JH, Lee SY, Tchah H. Mitomycin C-induced reduction of keratocytes and fibroblasts after photorefractive keratectomy. Invest Ophthalmol Vis Sci. 45, 2004, p. 2978-2984, Available at: http://www.iovs.org/content/45/9/2978.full.pdf
. Accessed April 1, 2013.
2. Santhiago MR, Netto MV, Wilson SE. Mitomycin C: biological effects and use in refractive surgery. Cornea
3. Carones F, Vigo L, Scandola E, Vacchini L. Evaluation of the prophylactic use of mitomycin-C to inhibit haze formation after photorefractive keratectomy. J Cataract Refract Surg
4. Hashemi H, Taheri SMR, Fotouhi A, Kheiltash A. Evaluation of the prophylactic use of mitomycin-C to inhibit haze formation after photorefractive keratectomy in high myopia: a prospective clinical study. BMC Ophthalmol. 4, 2004, 12, Available at: http://www.biomedcentral.com/content/pdf/1471-2415-4-12.pdf
. Accessed April 1, 2013.
5. Lee DH, Chung HS, Jeon YC, Boo SD, Yoon YD, Kim JG. Photorefractive keratectomy with intraoperative mitomycin-C application. J Cataract Refract Surg
6. Bedei A, Marabotti A, Giannecchini I, Ferretti C, Montagnani M, Martinucci C, Barabesi L. Photorefractive keratectomy in high myopic defects with or without intraoperative mitomycin C: 1-year results. Eur J Ophthalmol
7. Braunstein RE, Jain S, McCally RL, Stark WJ, Connolly PJ, Azar DT. Objective measurement of corneal light scattering after excimer laser keratectomy. Ophthalmology
8. Mφller-Pedersen T, Cavanagh HD, Petroll WM, Jester JV. Corneal haze development after PRK is regulated by volume of stromal tissue removal. Cornea
9. Kitazawa Y, Maekawa E, Sasaki S, Tokoro T, Mochizuki M, Ito S. Cooling effect on excimer laser photorefractive keratectomy. J Cataract Refract Surg
10. Maldonado-Codina C, Morgan PB, Efron N. Thermal consequences of photorefractive keratectomy. Cornea
11. Mohan RR, Hutcheon AEK, Choi R, Hong JW, Lee JS, Mohan RR, Ambrósio R Jr, Zieske JD, Wilson SE. Apoptosis, necrosis, proliferation, and myofibroblast generation in the stroma following LASIK and PRK. Exp Eye Res
12. Barbosa FL, Chaurasia SS, Kaur H, de Medeiros FW, Agrawal V, Wilson SE. Stromal interleukin-1 expression in the cornea after haze-associated injury. Exp Eye Res
13. Thomas KE, Brunstetter T, Rogers S, Sheridan MV. Astigmatism: risk factor for postoperative corneal haze in conventional myopic photorefractive keratectomy. J Cataract Refract Surg
14. Hefetz L, Domnitz Y, Haviv D, Krakowsky D, Kibarsky Y, Abrahami S, Nemet P. Influence of patient age on refraction and corneal haze after photorefractive keratectomy. Br J Ophthalmol. 81, 1997, p. 637-638, Available at: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1722288/pdf/v081p00637.pdf
. Accessed April 1, 2013.
15. Tanzer DJ, Isfahani A, Schallhorn SC, LaBree LD, McDonnell PJ. Photorefractive keratectomy in African Americans including those with known dermatologic keloid formation. Am J Ophthalmol
16. Tabbara KF, El-Sheikh HF, Sharara NA, Aabed B. Corneal haze among blue eyes and brown eyes after photorefractive keratectomy. Ophthalmology
17. Stojanovic A, Nitter TA. Correlation between ultraviolet radiation level and the incidence of late-onset corneal haze after photorefractive keratectomy. J Cataract Refract Surg
18. Thornton I, Xu M, Krueger RR. Comparison of standard (0.02%) and low dose (0.002%) mitomycin C in the prevention of corneal haze following surface ablation for myopia. J Refract Surg
19. Leccisotti A. Mitomycin-C in hyperopic photorefractive keratectomy. J Cataract Refract Surg
20. Virasch VV, Majmudar PA, Epstein RJ, Vaidya NS, Dennis RF. Reduced application time for prophylactic mitomycin C in photorefractive keratectomy. Ophthalmology
21. Pfister RR. Permanent corneal edema resulting from the treatment of PTK corneal haze with mitomycin; a case report. Cornea
22. Diakonis VF, Pallikaris A, Kymionis GD, Markomanolakis MM. Alterations in endothelial cell density after photorefractive keratectomy with adjuvant mitomycin. Am J Ophthalmol
23. Goldsberry DH, Epstein RJ, Majmudar PA, Epstein RH, Dennis RF, Holley G, Edelhauser HF. Effect of mitomycin C on the corneal endothelium when used for corneal subepithelial haze prophylaxis following photorefractive keratectomy. J Refract Surg
24. Morales AJ, Zadok D, Mora-Retana R, Martínez-Gama E, Robledo NE, Chayet AS. Intraoperative mitomycin and corneal endothelium after photorefractive keratectomy. Am J Ophthalmol
25. Nassiri N, Farahangiz S, Rahnavardi M, Rahmani L, Nassiri N. Corneal endothelial cell injury induced by mitomycin-C in photorefractive keratectomy: nonrandomized controlled trial. J Cataract Refract Surg
26. Wallau AD, Campos M. Photorefractive keratectomy with mitomycin C versus LASIK in custom surgeries for myopia: a bilateral prospective randomized clinical trial. J Refract Surg
27. Gambato C, Ghirlando A, Moretto E, Busato F, Midena E. Mitomycin C modulation of corneal wound healing after photorefractive keratectomy in highly myopic eyes. Ophthalmology
. 2005;112:208-218. discussion by RS Rubinfeld, 219.
28. Gartry DS, Kerr Muir MG, Lohmann CP, Marshall J. The effect of topical corticosteroids on refractive outcome and corneal haze after photorefractive keratectomy; a prospective, randomized, double-blind trial. Arch Ophthalmol
29. O’Brart DPS, Lohmann CP, Klonos G, Corbett MC, Pollock WST, Kerr-Muir MG, Marshall J. The effects of topical corticosteroids and plasmin inhibitors on refractive outcome, haze, and visual performance after photorefractive keratectomy; a prospective, randomized, observer-masked study. Ophthalmology