The Effects of Topical Corticosteroids and Plasmin Inhibitors on Refractive Outcome, Haze, and Visual Performance after Photorefractive Keratectomy

The Effects of Topical Corticosteroids and Plasmin Inhibitors on Refractive Outcome, Haze, and Visual Performance after Photorefractive Keratectomy A Prospective, Randomized, Observer,masked Study David P. S. O'Brart, FRCS, FRCOphth, 1 ChrisP. Lohmann, MD/· 2 Gregory Klonos, MD, 1 Melanie C. Corbett, FRCS, 1 William S. T. Pollock, FRCS, 1 Malcolm G. Kerr-Muir, FRCS, FRCOphth/ John Marshall, PhD1 Background: This study of 86 patients with 12 months of follow-up was designed to determine whether topical corticosteroids or plasmin inhibitors have an effect on the outcome of photorefractive keratectomy. Methods: Patients were allocated randomly to either steroid (0.1% fluorometholone for 6 months), plasmin-inhibitor (aprotinin 40 IUfml for 3 weeks), or control (no treatment) groups and underwent either -3.00- or -6.00-diopter (D) corrections. Results: With -3.00-D corrections, the mean refractive change was significantly greater at 3 and 6 months (P < 0.05) in the steroid group compared with the control group. When steroids were discontinued, the difference became insignificant within 3 months. Similarly, with -6.00-D procedures the mean refractive change was greater at 6 weeks and 3 and 6 months (P < 0.01 ), but the refractive change again became insignificant 3 months after stopping steroid treatment. Four patients treated with steroids had a hyperopic shift greater than +2.00 D of that intended at 12 months. Similar overcorrections were not noted in the other treatment groups. There were no differences in refractive outcome between the aprotinin and control groups at any stage. With -6.00-D procedures, objective measurements of haze were significantly greater in the aprotinin group compared with the control group at 9 and 12 months (P < 0.05). With this exception, there were no differences in haze, forward or backward scatter of light, best-corrected visual acuity, or halo measurements between the groups. Conclusions: Corticosteroids can maintain a hyperopic shift during their administration, but this effect is reversed on cessation of treatment. Objective tests have shown that steroids have no effect on corneal haze or visual performance after PRK. There is no justification for routinely submitting all patients to long-term steroid regimens and their associated side effects. Treatment with aprotinin produced no beneficial effect on refractive outcome, and haze was greater in the -6.00-D procedures. The concept of modulating the plasminogen activator/plasmin system to regulate wound healing after PRK is discussed. Ophthalmology 1994; 101:1565-1574

Originally received: November 17, 1993. Revision accepted: March 28, 1994. 1

Department of Ophthalmology, St. Thomas' Hospital, London.

2

University Eye Clinic, Regensberg, Germany.

Supported in part by equipment funding and a research fellowship sponsored by the Iris Fund for Prevention of Blindness (Dr. O'Brart) and a grant from the Commission of the European Communities

(Dr. Klonos), London, UK. Dr. Corbett holds the William's Fellowship for Medical and Scientific Research of the University of London. Professor Marshall is a consultant for Summit Technology. Reprint requests to David P. S. O'Brart, FRCS, FRCOphth, United Medical and Dental Schools, Department of Ophthalmology, St. Thomas' Hospital, London, England SE I 7EH.

1565

Ophthalmology

Volume 101, Number 9, September 1994

Excimer laser photorefractive keratectomy (PRK) has been the subject of clinical trials for 4 years. More than 80% of eyes with corrections of -6.00 diopters (D) or less have a refractive outcome of± 1.00 D of that intended. 1-4 The refractive change stabilizes a few months after surgery, and complications impairing postoperative visual performance are few. In all published studies, the predictability and efficacy of the procedure tend to decrease with increasing dioptric corrections. 1- 4 However, there is also some variation in outcome between individuals with lower diopter procedures. 1 After PRK, highly complex wound healing mechanisms are invoked, which directly influence the outcome of the procedure. Durrie (personal communication) retrospectively has assigned patients into three classes of wound healing responses based on their refractive outcome. The majority he describes as "normal," and these individuals showed only mild, transient disturbances in corneal transparency (haze), together with a near-emmetropic outcome. A small group of his patients he describes as having "poor" wound healing responses with clear corneas throughout the postoperative period and persistent hyperopia. The remainder had an aggressive response with marked haze and regression of the intended correction. As yet, techniques have not been devised to identify these wound healing types before surgery. Wound healing in the cornea involves three mechanisms: the removal of debris and damaged tissue, the repair of damaged structures, and the replacement of irreparable systems. The relative contribution of each of these mechanisms varies between the epithelium and stroma. In the epithelium, replacement is the predominant mechanism. In the stroma, the wound healing processes are regulated by two systems: the plasminogen-activator/ plasmin system, which initiates the removal of damaged tissue, and the activated keratocyte system, which is involved in the replacement of damaged tissue by the synthesis of collagen and glycosaminoglycans. There has been considerable interest in the pharmacologic control of these healing processes to improve further the predictability of PRK. Most research has been directed at inhibiting the activated keratocyte system with topical corticosteroids. Studies have indicated that these agents can modify corneal reparative mechanisms after conventional surgical injury. A reduction in the tensile strength of incisions in rabbit cornea has been demonstrated after their use. 5-7 The mechanism of action has not been elucidated fully, but it has been suggested that steroids interfere with DNA synthesis in keratocytes, producing diminished cellular activity and reduced collagen synthesis. 8- 10 In a limited study in which PRK was performed on six rabbits, Tuft et al 11 reported that haze was reduced in steroid-treated corneas compared with control corneas. Although the authors did not examine refractive changes, they suggested that wound healing after PRK might be modulated by these agents. As a result of this limited evidence, all groups performing early clinical trials of PRK used postoperative topical steroid regimens. 1- 4 A variety of agents was used, together with varying dose regimens. Different groups attributed

1566

different degrees of clinical importance to the use of steroids, with some suggesting that the frequency of administration of these agents could be titrated directly against the postoperative progress of haze and regression. 2 Other studies suggested a more limited role, with reports of no discernible differences in outcome between patients who defaulted on the steroid regimen and those who were compliant. 1 To address these issues, Gartry et al 12 performed a prospective, randomized, double-blind trial. This indicated that topical corticosteroids had a significant effect on refractive outcome during the first 3 months after PRK. However, after 3 months, at which time steroids were discontinued, no statistically significant difference could be demonstrated. Haze was not influenced by the use of steroids at any stage postoperatively. Gartry et al concluded that, in view of the patients' side effects, it would be unacceptable to use steroids to maintain only a temporary benefit on refraction. Despite the evidence presented by this study, the majority of ophthalmologists still prescribe steroids after PRK. A number of criticisms have been proffered to challenge the results of the Gartry et al 12 study. These have included their use of small ablation zones (4.00 mm), the high incidence of regressions among their cohort of patients, and the duration of steroid treatment. Conflicting results also have been reported, 13• 14 although these studies are retrospective, nonrandomized, unmasked, and based on small numbers of patients. Laboratory studies have indicated that the major phases of wound healing occur within the first few months after surgery. However, immunohistochemical studies of excimer laser-ablated monkey corneas have demonstrated that some limited remodeling is evident as late as 18 months after photoablation. 15 In addition, some anecdotal case reports have claimed that reintroduction of topical corticosteroids many months after surgery may reverse regression and improve haze. 16 • 17 Such observations may indicate that although the replacement phases of wound healing should have been completed by 3 months postoperatively, the longer-term use of steroids may have a beneficial effect. There have been no clinical studies examining the effects of 6-month corticosteroid regimens. To address these issues, we have performed a prospective, randomized, observer-masked study in which some patients were treated with 0.1% fluorometholone for 6 months. The concept of modulating the plasminogen-activator/ plasmin system to regulate wound healing after PRK has been suggested by Lohmann and Marshall. 18 Plasmin is a serine protease derived from plasminogen, which is a normal component of tears. 19 Plasmin activates several enzymes such as procollagenase and macrophage elastase and degrades many matrix proteins such as fibronectin and laminin. 2 Fibronectin is an adhesive cell-surface glycoprotein that mediates cross-linking of collagen and is important in epithelial cell migration. The concentration of plasmin in the tears of healthy eyes usually is below 0.2 ~tg/ml. This concentration is known to increase with ocular surface pathology, 21 •22 ex-

°

O'Brart et al · Topical Corticosteroids and Plasmin Inhibitors after PRK posure to allergens, 23 contact lens use, 24 anterior keratectomy, 19 and immediately after PRK. 25 In addition, it has been suggested that high preoperative concentrations may occur in patients with a propensity to an aggressive wound healing response. 25 Aprotinin is a polypeptide extracted from bovine lung tissue, which inhibits plasmin. It has been used in animal studies and clinical trials for various corneal disorders, such as chronic ulcers, and the results have been very encouraging. 21 Ideally, after PRK, there should be removal of surface debris, followed by epithelial cell migration and replacement. In the stroma, removal and replacement should be minimal, and damaged tissues should be repaired. It has been postulated that by inhibiting the actions of plasmin, epithelial regeneration might be facilitated and stromal removal limited, both of which are desirable features. To examine these concepts, we performed a prospective, randomized, observer-masked study in which some patients were treated with aprotinin.

Subjects and Methods Subjects The study comprised 86 patients (86 eyes) who underwent PRK for correction of myopia. The mean age was 37 years (range, 24-68 years). To minimize interpatient variation due to different ablation depths resulting from a variety of different corrections, only -3.00- or -6.00D procedures were undertaken (programmed ablation depths of 36 and 62 ~tm, respectively). We therefore selected patients with preoperative refractions close to -3.00 D (mean, -3.04 D; range, -2.5 to -3.625 D) and -6.00 D (mean, 6.14 D; range, -5.5 to -7.0 D). There were 47 patients who underwent - 3.00-D procedures and 39 who underwent -6.00-D corrections.

Patient Assessment After obtaining Ethical Committee approval, prospective patients were sent background information and a questionnaire. Those considered suitable for the study were interviewed and fully counseled before PRK to discuss the investigative nature of the study. Past optometric records were scrutinized to ensure stability of refraction, and all subjects were older than 24 years of age. All eyes had less than 1.50 D of refractive astigmatism. Patients with pre-existing ocular pathology, diabetes, or connective tissue disorders were excluded. Preoperatively, a detailed ocular examination was performed, including refraction, keratometry, biomicroscopy, tonometry, and mydriatic funduscopy. Objective measurements of disturbances in corneal transparency were made with a charge-coupled device camera system, the calibration and data analysis of which have been described in detail previously. 26 In essence, this device measured gray-scale disturbance caused by the combined signal of light reflected and scattered back from the cornea or backscattered light alone. A charge-coupled device camera

mounted on a Haag-Streit slit lamp was used to capture an image of the cornea on a frame grabber. The image then was digitized and analyzed using inhouse software. To discriminate between reflected and scattered light, linear polarizing filters within the charge-coupled device camera and slit-lamp light source were used.

Procedure The operative procedure has been described elsewhere. 1 All patients were treated by a single surgeon, during a 6week period in May and June 1992. A Summit Technology ExciMed UV200 excimer laser (Boston, MA) was used, with an emission wavelength of 193 nm, a fixed pulse repetition rate of 10 Hz, and a radiant exposure of 180 mJ/cm 2• In this system, differential ablation is achieved by successive laser pulses delivered through an expanding iris diaphragm. The maximum diameter of the diaphragm was 5.00 mm in all patients.

Postoperative Treatment Patients were allocated by means of a random number system into the following three treatment groups: 1. Group 1 received 0.1% fluorometholone, one drop immediately postoperatively. After removal ofthe eye pad the next morning, drops then were administered twice hourly for the first week and 8 times daily for the first month. The dosage was reduced to six times daily for the second month, five times daily for the third, four times a day for the fourth, three times daily for the fifth, twice daily for 3 weeks, and once daily for 2 weeks. Fluorometholone was used because it is prescribed routinely after PRK in a large number of centers, and it is claimed to be efficacious. 4 It was ideally suited for the current study in view of the long duration of treatment. Fluorometholone is associated with a lower incidence of steroid-induced raised intraocular pressure, 27 compared with other topical corticosteroids and has limited penetration into the aqueous humor.28 2. Group 2 received aprotinin 40 IU/ml, one drop immediately postoperatively and after removal of the eye pad, five times daily for 3 weeks. Treatment was limited to 3 weeks because aprotinin is a large 58amino acid polypeptide that cannot penetrate the normal corneal epithelial barrier; hence, treatment after this time would be ineffective. 3. Group 3 was a "control" group and received no corticosteroid or plasmin-inhibitors postoperatively. In all patients, 1% atropine eyedrops and 1% chloramphenicol ointment were applied immediately after the procedure. Oral analgesics were prescribed, and then the eye was padded lightly overnight. Chloramphenicol eyedrops were administered four times daily from the time of removing the pad for 2 weeks postoperatively.

1567

Ophthalmology

Volume 101, Number 9, September 1994

Postoperative Assessment Postoperative examinations were performed at l and 6 weeks and 3, 6, 9, and 12 months. At each visit, refraction and slit-lamp examination were performed. Objective measurements of anterior stromal haze and back-scattered light were made at each visit. 26 The degradation of the retinal image generated by the forward scatter oflight was assessed using a recently described computerized technique.29 In summary, this was a two-part test in which visual contrast was measured first with central test stimulus generated on a high-resolution monitor. This stimulus flickered at 7.5 Hz between the background level and a match level. The patient adjusted the contrast between the match luminance and background to minimize flicker by means of buttons on a computer keyboard. The test then was repeated with a bright, annular light source flickering in counterphase and surrounding the central test stimulus. This "straylight" provided an additional luminance for the forward scatter of light. At 12 months, one additional test was performed, which measured the area of the myopic blur circle that was generated around a central bright light source on a high-resolution monitor. 30 Although patients were aware of whether they were instilling drops postoperatively; the investigators performing the postoperative examinations were not. The study, therefore, was "observer-masked."

Statistical Methods The unpaired Student's t test was used to compare the mean changes in refraction, objective haze measurements, and light scatter at each postoperative visit. Wilcoxon's rank-sum tests were used to compare measurements of halo because these were not distributed normally. Results with P < 0.05 were considered statistically significant.

Results Refractive Outcome The mean changes in refraction with time for patients treated with and without steroids are shown in Figures 1 and 2 for the -3.00- and -6.00-D corrections, respectively. In all groups, there was a characteristic hyperopic shift during the first few weeks postoperatively, which regressed toward emmetropia within the first 3 months. In those patients who underwent - 3.00-D corrections, the mean change in refraction was significantly greater at 3 and 6 months (P < 0.05) in the steroid group. However, within 3 months of steroids being discontinued, this difference became insignificant (P > 0.1 ). The 95% confidence interval at 12 months was -0.43 to 0.91 D. In -6.00-D groups, the mean change in refraction was significantly greater at 6 weeks and at 3 and 6 months (P < 0.01) in the steroid-treated patients. However, within 3 months of discontinuing steroids this difference again fell below statistical significance (P < 0.07). In this group

1568

the 95% confidence interval was -0.03 to 2.6 D at 12 months. In the steroid group, four patients (14%) retained a hyperopic shift greater than +2.00 D of that intended at 12 months. Similar overcorrections were not noted in the other groups. None of the overcorrected patients was found to have an intraocular pressure greater than 20 mmHg at any stage postoperatively. There were no refractive differences between the aprotinin and control groups at any stage for either -3.00- or -6.00-D procedures (Figs 3 and 4).

Objective Haze Measurements The results of objective measurements of anterior stromal haze for reflected and scattered light and back-scattered light alone are shown for the steroid and control groups in Figures 5 and 6. There was no statistically significant difference in the amount of haze or back scatter of light between the steroid and control groups at any stage. Similar results for the aprotinin and control eyes are shown in Figures 7 and 8. In the -6.00-D corrections treated with aprotinin, objective measurements of thereflected and scattered light signal from corneal haze were greater at all stages postoperatively and reached statistically significant levels at 9 and 12 months (P < 0.05). Importantly, there were no significant differences in the measurements of the back-scattered light signal alone. In all groups, haze reached a maximum at 3 months in the - 3.00-D eyes and at 6 months with the -6.00-D corrections. The haze declined thereafter. The degree of haze and back scatter of light was significantly greater in the -6.00- compared with the - 3.00-D corrections at all stages postoperatively (P < 0.01).

--

Cl

-... ~~ c: 0

5

p
p
()

ctl

4

( I)

----~ ----- - - -

0:

·= (I)

Ol

c: ctl

..c ()

0

01

6

13

26

39

52

Weeks After PRK. Figure 1. The mean change in refraction with time for eyes treated with steroids (triangles) (n = 16), compared with control eyes (squares) (n = 16), and undergoing -3.00-diopter corrections. P values at which the differences between the two groups reached statistical significance are shown.

O'Brart et al · Topical Corticosteroids and Plasmin Inhibitors after PRK 12 ~--~----~--------~--------~---------+

9.

10

- ~~ c::

~p<0.01

9.

p<0.01

t\ """'

0

( .)

p<0.07

--~-------

«<

..... Q)

a:

Q) 0)

------------ _______

·-----..-_

~r-----~----~----~

4

10

c::

p<0.01

0

-

8

·.;:; (.)

«<

..... Q)

a:

.5

4

Q) 0)

c::

c::

«< ..c:

«< ..c:

2

()

()

0 ~---r----r--------.--------,---------r 13 39 26 52 6 01

01

6

13

52

Weeks After PRK.

Weeks After PRK. Figure 2. The mean change in refraction with time for eyes treated with steroids (triangles) (n = 13), compared with controls (squares) (n = 14), and undergoing -6.00-diopter corrections. P values at which the differences between the two groups reached statistical significance are shown.

Forward Light Scatter Measurements of visual contrast sensitivity with and without the straylight source are shown in Figures 7 and 8. Disturbances were maximal at 3 months and improved thereafter. There were no statistically significant differences between steroid- and aprotinin-treated patients and control subjects at any stage postoperatively. There were no differences between the -3.00- or -6.00-D procedures.

Halo Measurement Computerized measurements of halo 12 months after PRK are listed in Table 1. The magnitude of halo was

Figure 4. The mean change in refraction with time for eyes treated with aprotinin (stars) (n = 12), compared with controls (squares) (n = 14), and undergoing -6.00-diopter corrections.

significantly greater in eyes that had received -6.00compared with - 3.00-D corrections (P < 0.05). There were no differences between the three treatment groups.

Complications A list of complications between patients in the three treatment groups is given in Table 2. There were no obvious differences between the three groups, except that transient intraocular pressure rises were recorded in two patients treated with steroids. The excessive hyperopic shifts in four eyes treated with steroids have been described in the Refractive Outcome section.

-

-

-

0

V)

240

'2 ::>

220

Q)

«< 200

c:: 0 ·.;:;

Reflected and Scattered Light

(.)

(/)

- ~~ (.)

«< .....

39

26

4

>-

180

..... G

-

160

N

140

Q)

Q)

a:

-

.5

---

Q)

«<

Q) 0)

::c

«<

>

-

c::

Q)

..c:

()

(.) Q)

.0

0 01

13

26

39

52

Weeks After PRK. Figure 3. The mean change in refraction with time for eyes treated with aprotinin (stars) (n = 15), compared with controls (squares) (n = 16), and undergoing - 3.00-diopter corrections.

0

120 100

Scattered Light Alone

80 0

13

26

39

52

Weeks After PRK Figure 5. Objective measurements of anterior stromal haze for eyes treated with steroids (triangles) (n = 16), compared with controls (squares) (n = 16), and undergoing - 3.00-diopter corrections.

1569

Ophthalmology

-

-

(/)

-

240

(/)

:::>

220

(13

(13

200

(,)

(,)

>

180

C!J

160

(!) ....

>-

(!) ....

CJ (!)

(!)

N

140

N

(13

-

(13

:c

-

120

(!)

Scattered Light Alone

100

(,) (!)

.~

..0

0

180 160 140 120

>

>

( ,)

200

(/)

(/)

(!)

p<0.05 220

(!)

~

:c

240

s:::

Reflected and Scattered Light

s:::

:::>

Volume 101, Number 9, September 1994

..0

80 0

13

26

39

52

0

Scattered Light Alone

100 80 0

13

26

39

52

Weeks After PRK.

Weeks After PRK. Figure 6. Objective measurements of anterior stromal haze for eyes treated with steroids (triangles) (n = 13), compared with controls (squares) (n = 14), and undergoing -6.00-diopter corrections.

Figure 8. Objective measurements of anterior stromal haze for eyes treated with aprotinin (stars) (n = 12), compared with controls (squares) (n = 14), and undergoing -6.00-diopter corrections. P values at which the differences between the two groups reached statistical significance are shown.

Discussion The Role of Topical Corticosteroids after Photorefractive Keratectomy The role of corticosteroids in optimizing wound healing after PRK is controversial. Histologic studies have documented the healing processes after photoablation in rabbit3 132 and monkey corneas. 33- 36 In the epithelium, regeneration occurs within a few days, and epithelial hyperplasia settles after several weeks. In the stroma, damaged keratocytes either degenerate or migrate out of the wound region. This initial phase is followed by repopu-

-

(i)

.'!:::

220

c:

:::>

Reflected and Scattered Light

200

(!) (13

(,)

(/)

>

.... (!)

180

160

C!J

G> 140 N (13

:c

120

G>

>

·;:;

100

G> ..0

80

(,)

0

Scattered Light Alone

0

13

26

39

52

Weeks After PRK. Figure 7. Objective measurements of anterior stromal haze for eyes treated with aprotinin (stars) (n = 15), compared with controls (squares) (n = 16), and undergoing - 3.00-diopter corrections.

1570

lation, because fibroblasts migrate into the edges of the wound area and lay down new collagen and extracellular matrix. During this period, subepithelial haze appears and becomes maximal at 3 to 4 months. 37 After 4 months, the number of activated keratocytes declines, and haze gradually clears. Limited studies in animal models using incisional wounds have demonstrated that topical steroids can modify corneal reparative mechanisms and decrease wound strength. 5-7 The timing of such treatment appears to be critical. Reduced wound strength is demonstrated when steroids are started immediately, but not if given after the sixth postoperative day. 5•6 However, surgical incisions have a very different aspect ratio from the wide excisional wounds ofPRK, and it cannot be assumed that steroids will have the same influence on healing. Tuft et al 12 performed 40-~tm photoablations in six rabbit corneas and reported less haze and a reduction of collagen synthesis in steroid-treated eyes compared with control eyes. Talamo et al 38 also reported a reduction in haze associated with steroid treatment, which was found to be enhanced with mitomycin C. However, the rabbit has no Bowman's layer, and the relative contributions of collagen types differ from the human cornea. In addition, all these animal studies were conducted on young animals whose eyes were still growing, whereas all potential patients have stable refractions with nongrowing eyes. The plasticity and regenerative capacity is likely to be greater in young, growing eyes. It is therefore not possible to extrapolate directly the wound healing responses from animal models to humans. All groups performing early clinical trials used corticosteroids postoperatively, although regimens differed. 1-4 Individual authors attributed varying degrees of clinical importance to the use of steroids, although none of these studies were controlled or designed to test the efficacy of

O'Brart et al · Topical Corticosteroids and Plasmin Inhibitors after PRK Table 1. Measurements of the Area of Halo Around a Bright Light Source Generated on a High-resolution Monitor30 in all 86 Eyes at 12 Months Postoperatively* Halo Area (mm2 ) -6.0D PRK

-3.0 D PRK Steroid

1412 1236 208-4907

Mean Median Range D

=

diopter; PRK

=

Aprotinin

Control

2001 1824 386-5114

1460 1338 568-3395

Steroid

Aprotinin

Control

3481 2542 1209-7712

2182 1879 470-4212

2794 2821 556-6080

photorefractive keratectomy.

• The mean value for 100 healthy myopic eyes corrected with spectacles (up to -7.5 D) was 444 mm2 (range, 0-1608 mm2).

these agents. Gartry et al 1 reported a series of 16 blind eyes in which no steroids were used. Haze was less than in their cohort of 120 sighted eyes that received 0.1% dexamethasone. In addition, considerable haze, regression, and individual variation were reported in the latter group, despite steroid treatment. 1 Of the 120 patients, 21 defaulted on the steroid regimen. There was no discernible difference in haze or regression between these patients and the remainder of the cohort. In contrast, Seiler and Wollensak2 found noncompliance to be a risk factor in the initiation of aggressive wound healing and suggested that the frequency of administration of steroid drops could be titrated against the postoperative progress of haze and regression. When designing any therapeutic regimen, risk versus benefit analysis must be undertaken. The capacity oftopical steroids to raise intraocular pressure, 39 reactivate herpes simplex40 and induce cataracts41 •42 is well documented. After PRK, the reported incidence of steroidinduced raised intraocular pressure varies between 10% 1 and more than 20%. 2 In the vast majority of patients, the incidence falls to normal levels after cessation of steroid treatment. Therapeutic agents rarely are required but

topical beta-blockers are effective. Currently, follow-up is limited to 4 years, and it is therefore not possible to measure the risk of cataracts developing. However, some investigators have reported cataracts developing after PRK, associated with long-term steroid treatment (Seiler T, personal communication). In the current study, 0.1% fluorometholone was used, and the incidence of raised intraocular pressure was only 7%. In view of the diverse opinions and potential side effects of steroids, Gartry et al 12 conducted a double-blind, placebo-controlled trial to examine the effects of these agents on the outcome of PRK. This trial demonstrated that steroids had no beneficial effect on refractive outcome after cessation of treatment, and haze remained unaffected at all stages postoperatively. However, despite its thorough design, the study was criticized for its high incidence of regressions and large standard deviations. 13 Clinical reports conflicting Gartry et al's results have since been published, 13 • 14 although these investigations are largely retrospective, nonrandomized, unmasked, and report small series of patients. The current study supports Gartry et al's observations, in that we found significantly less regression of refraction during steroid administration, but

Table 2. Complications of Photorefractive Keratetomy at 12 Months -6.0DPRK

-3.0DPRK Complications

Steroid

Aprotinin

Control

Steroid

Aprotinin

Control

Total%*

Hyperopic shift > 2.0 D of intended Reduced BCVA < 0.5 line Disturbances of night vision Severe night vision impairmentt Raised intraocular pressure Symptoms of epithelial instabilityt Tenderness

1 2 4 1 1 0 4

0 2 5 1 0 1 5

0 1 5 1 0 1 6

3 1 6 2 1 0 5

0 1 6 0 0 0 3

0 2 6 0 0 1 3

5%(14%) 10% 37% 6% 2(7%) 3.5% 30%

D = diopter; PRK = photorefractive keratectomy; BCVA = best-corrected visual acuity. • Values in parentheses indicate the steroid-treated eyes alone.

t Defined as an inability to drive at night with the treated eye alone. 1' All eyes settled within 3 months of treatment.

1571

Volume 101, Number 9, September 1994

Ophthalmology

.... ....c... (I)

30

CIJ

0 (.) ~

...

20

CIJ

15

-....

Straylight Source On

25

Q)

()

(/)

....

..::::

10

0>

_J

....

"'C

Straylight Source Off

CIJ

....~ 0

u..

0 0

13

26

39

52

Weeks After PRK. Figure 9. Measurements of visual contrast sensitivity with and without the flickering straylight source for eyes undergoing - 3.00-diopter corrections in all three treatment groups. The normal mean values for healthy myopic eyes with best-spectacle correction (up to -7.50 diopters) is (1) without straylight source (n = 100), 2.97% contrast (range, 0%-6.6%), and (2) with straylight source (n = 100), 9.1% contrast (range, 1.7%18.9%). Stars = eyes treated with aprotinin; squares = controls; triangles = eyes treated with steroids.

this was not maintained 3 months after cessation of treatment. Concern may be expressed over the small sample sizes in the current study. In the -6.00-D correction group, the confidence limits suggest a trend that steroids are having an effect on refractive outcome even after cessation of treatment. In the- 3.00-D group, the confidence limits do not suggest such a trend. For high-dioptric corrections, further studies are required with larger group sizes. However, there were no differences in haze, forward or backward scatter of light, best-corrected visual acuity, or halo measurements between the steroid and control groups. Although the numbers are small, the increased tendency of sustained hyperopic shifts in some eyes treated with steroids is a cause for concern. We believe there is no justification for subjecting all patients to long-term steroid regimens and their associated side effects. As databases increase, it will be determined whether an alteration in mean refractive change is necessary. If this is the case, it can be achieved by alteration of the algorithm controlling the laser. A number of recent studies have reported a significant reversal of regression and improvement in haze after reintroduction of steroid treatment. 16 • 17 However, these studies are based on anecdotal evidence from retrospective analysis of small series of patients with limited follow-up, and further investigations are required. The time course of re-administration of steroids postoperatively and the relations to the time course of wound healing are particularly confusing. Some groups report major changes in refraction with steroid treatment over 6 months after PRK and with resultant changes being affected within hours or days. The possible mechanisms for corticosteroid action

1572

in these eyes are unclear. Given the lapse of time after surgery, there would be no active collagen synthesis and therefore interference with this process seems unlikely to be responsible for the reported results. In addition, interference with collagen synthesis would not produce such rapid changes. An effect on stromal hydration could be postulated. Steroids may act via hyaluronic acid, which is a high molecular weight disaccharide polymer with the capacity to bind large amounts of water. Increased amounts of hyaluronic acid are known to be present in rabbit corneas after PRK43 and can be reduced with topical corticosteroids (Fitzsimmons, unpublished data; presented at the 1992 Association for Research in Vision and Ophthalmology Annual Meeting). However, in animal studies this reduction of hyaluronic acid has only been achieved by steroid administration during the "active" phase of wound healing. The role of steroids after PRK appears to be limited. There is no justification for their routine administration in view of their minimal, largely transient, effects on refractive correction and their potential side effects. If corticosteroids are to have a role in improving the outcome of PRK, it may be in selected patients with aggressive healing responses. Further studies are required to ascertain their capacity to reverse regression and haze in these eyes.

Modulation of the Plasminogen-activator/ Plasmin System after Photorefractive Keratectomy The wound created by the excimer laser is unique, in that tissue is ablated with minimal damage to adjacent struc-

.... ........c (I)

30

CIJ

0

(.) ~

....

20

CIJ

15

........

Straylight Source On

25

Q)

()

(/)

..::::

10

0>

_J

... ...0

"'C CIJ

~

u..

0 0

13

26

39

52

Weeks After PRK. Figure 10. Measurements of visual contrast sensitivity with and without the flickering straylight source for eyes undergoing -6.00-diopter corrections in all three treatment groups. The normal mean values for healthy myopic eyes with best-spectacle correction (up to -7.50 diopters) is (1) without straylight source (n = 100), 2.97% contrast (range, 0%-6.6%), and (2) with straylight source (n = 100), 9.1% contrast (range, 1.7%18.9%). Stars = eyes treated with aprotinin; squares = controls; triangles = eyes treated with steroids.

O'Brart et al · Topical Corticosteroids and Plasmin Inhibitors after PRK tures. 44 However, even this minute amount of damaged tissue is sufficient to activate and release the factors responsible for wound healing. The release of these chemical mediators, proteolytic enzymes, etc., leads to an enlargement of the wound area and further involvement of the adjacent tissues. Thus, the wound healing process amplifies the area of damage and will result in a larger wound than initiated by the excimer laser. The concept of modulating the plasminogen-activator/ plasmin system to regulate wound healing after PRK was suggested recently by Lohmann and Marshall. 18 The inactive precursor plasminogen is activated to plasmin by plasminogen activators. The two types of plasminogen activators are ( l) tissue type and (2) urokinase type. After wounding in rabbit corneas, an increase in urokinase-type plasminogen activator has been demonstrated in the stroma immediately beneath the wound bed and at the wound edges. 45 Lohmann and Marshall 18 used immunohistochemical techniques to investigate the presence of plasminogen activators after excimer laser PRK. They have found strong immunoreactivity in the epithelium overlying the ablated area and a weak staining in the anterior stroma from 3 days to 3 months postoperatively. Plasmin degrades matrix proteins such as fibronectin and laminin and activates enzymes such as procollagenase and macrophage elastase. Fibronectin and laminin are major factors in the extracellular matrix 20 and play an important role in the adhesion of both the epithelium to its underlying basement membrane and the basement membrane to the underlying stroma. Fibronectin is important in intercellular associations and is thought to function as an adhesive cell-surface glycoprotein. This is of particular importance in corneal wound healing because epithelial cell migration will involve rapid and repeated synthesis and degradation of surface adhesive elements. Lohmann et al 25 have found elevated plasmin concentrations in tear samples taken immediately after PRK. They also found high levels (.s351lg/ml) preoperatively in three patients who all showed early regression (within 6 weeks) of their PRK corrections.2 5 This may indicate that high preoperative levels of plasmin may be associated with aggressive wound healers. We found no significant beneficial effect in refractive outcome and visual performance in eyes treated with aprotinin compared with the control eyes. In the -6.00D corrections group, objective measurements of the reflected and scattered light signal from disturbances in corneal transparency were actually greater at all stages postoperatively and reached statistically significant levels at 9 and 12 months in aprotinin-treated eyes. There were no differences in measurements of back-scattered light alone, which have been shown to have a greater correlation with low-contrast visual acuity loss. 46 This is supported by the forward-scattered light measurements, which showed no differences between the groups (Figs 9 and 10). These haze results, although disappointing, do demonstrate that it is possible to modify wound healing via the plasminogen activator/plasmin system and produce statistically significant effects on the outcome ofPRK. This is of importance because aprotinin was given only in low dose for 3 weeks,

and its effects persisted for many months after treatment. In contrast, we have not been able to demonstrate any permanent changes with long-term corticosteroid therapy using similar group sizes. It can be postulated that by inhibiting only the removal phase, an imbalance in the wound healing process was created, which resulted in the increased deposition of new collagen and glycosaminoglycans at the wound site. By inhibiting both the removal and replacement phases of healing, an optimal situation may be obtained. Studies using a combination of aprotinin and corticosteroids after PRK currently are under way.

References I. Gartry DS, Kerr Muir MG, Marshall J. Photorefractive keratectomy with an argon fluoride excimer laser: a clinical study. Refract Corneal Surg 1991 ;7:420-35. 2. Seiler T, Wollensak J. Myopic photorefractive keratectomy with the excimer laser. One-year follow-up. Ophthalmology 1991 ;98: 1156-63. 3. McDonald MB, Liu JC, Byrd TJ, et al. Central photorefractive keratectomy for myopia: partially sighted and normally sighted eyes. Ophthalmology 1991;98:1327-37. 4. Sher NA, Chen V, Bowers RA, et al. The use of the 193nm excimer laser for myopic photorefractive keratectomy in sighted eyes: a multicenter study. Arch Ophthalmol 1991; I 09: 1525-30. 5. Sugar J, Chandler JW. Experimental corneal wound strength: effect of topically applied corticosteroids. Arch Ophthalmol 1974;92:248-9. 6. Phillip K, Arffa R, Cintron C, et al. Effect of prednisolone and medroxyprogesterone on corneal wound healing, ulceration and neovascu1arization. Arch Ophthalmol 1983; 101:640-3. 7. Newell FW, Dixon JM. Effect of subconjunctival cortisone upon the immediate union of experimental corneal grafts. Am J Ophthalmol 1951;34:977-81. 8. Polack FM, Rosen PN. Topical steroids and tritiated thymidine uptake: effect on corneal healing. Arch Ophthalmol 1967;77:400-4. 9. Gasset AR, Lorenzetti DWC, Ellison EM, Kaufman HE. Quantitative corticosteroid effect on corneal wound healing. Arch Ophthalmol 1969;81 :589-91. 10. McDonald TO, Borgmann AR, Roberts MD, Fox LG. Corneal wound healing. I: Inhibition of stromal healing by three dexamethasone derivatives. Invest Ophthalmol 1970;9:7039. 11. Tuft SJ, Zabel RW, Marshall J. Corneal repair following keratectomy: a comparison between conventional surgery and laser photoablation. Invest Ophthalmol Vis Sci 1989;30: 1769-77. 12. 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 1992;110:944-52. 13. Tengroth B, Fagerholm P, SOderberg P, et al. Effect of corticosteroids in postoperative care following photorefractive keratectomies. Refract Corneal Surg 1993;9(Suppl):S61-4. 14. Tengroth B, Epstein D, Fagerholm P, eta!. Excimer laser photorefractive keratectomy for myopia: clinical results in sighted eyes. Ophthalmology 1993;100:739-45.

1573

Ophthalmology

Volume 101, Number 9, September 1994

15. SundarRaj N, Geiss MJ III, Fantes F, et al. Healing of excimer laser ablated monkey corneas: an immunohistochemical evaluation. Arch Ophthalmol 1990; 108: 1604-10. 16. Carones F, Brancato R, Venturi E, eta!. Efficacy of corticosteroids in reversing regression after myopic photorefractive keratectomy. Refract Corneal Surg !993;9(Suppl):S52-6. 17. Fitzsimmons TD, Fagerholm P, Tengroth B. Steroid treatment of myopic regression: acute refractive and topographic changes in excimer photorefractive keratectomy patients. Cornea 1993; 12:358-61. 18. Lohmann CP, Marshall J. Plasmin- and plasminogen-activator inhibitors after excimer laser photorefractive keratectomy: new concept in prevention of postoperative myopic regression and haze. Refract Corneal Surg 1993;9:300-2. 19. van Setten G-B, Salonen E-M, Vaheri A, eta!. Plasmin and plasminogen activator activities in tear fluid during corneal wound healing after anterior keratectomy. Curr Eye Res 1989;8: 1293-8. 20. Ding M, Burstein NL. Review: fibronectin in corneal wound healing J Ocul Pharmacol 1988;4:75-91. 21. Salonen E-M, Tervo T, Torma E, et al. Plasmin in tear fluid of patients with corneal ulcers: basis for a new therapy. Acta Ophthalmol 1987;65:3-12. 22. Tervo T, Salonen E-M, Vahen A, eta!. Elevation of tear fluid plasmin in corneal disease. Acta Ophthalmol 1988;66: 393-99. 23. Salonen E-M, Lauharanta J, Sim PS, et al. Rapid appearance of plasmin in tear fluid after ocular allergen exposure. Clin Exp Immunol 1988;73:146-8. 24. Tervo T, van Setten G-B, Andersson R, eta!. Contact lens wear is associated with the appearance of plasmin in the tear fluid-preliminary results. Graefes Arch Clin Exp Ophthalmol 1989;227:42-4. 25. Lohmann CP, O'Brart DPS, Patmore A, et al. Plasmin in the tear fluid: a new therapeutic concept to reduce postoperative myopic regression and corneal haze after excimer laser photorefractive keratectomy. Lasers Light Ophthalmol 1993;5:205-10. 26. Lohmann CP, Timberlake GT, Fitzke FW, et al. Corneal light scattering after excimer laser photorefractive keratectomy: the objective measurements of haze. Refract Corneal Surg 1992;8: 114-21. 27. Fairbairn WD, Thorson JC. Arch Ophthalmol1971 ;86: 138-41. 28. McGhee CNJ, Watson DG, Midgley JM, eta!. Penetration of synthetic corticosteroids into the aqueous humour. Eye 1990;4:526-30. 29. Lohmann CP, Fizke F, O'Brart D, eta!. Corneal light scattering and visual performance in myopic individuals with spectacles, contact lenses, or excimer laser photorefractive keratectomy. Am J Ophthalmol 1993;115:444-53. 30. O'Brart DPS, Lohmann CP, Fitzke FW, et al. Night vision disturbances after excimer laser photorefractive keratectomy: haze and halos. Eur J Ophthalmol 1994;4:43-51.

1574

31. TuftS, Marshall J, Rothery S. Stromal remodelling following photorefractive keratectomy. Lasers Ophthalmol 1987;1: 177-83. 32. Goodman GL, Trokel SL, Stark WJ, et al. Corneal healing following laser refractive keratectomy. Arch Ophthalmol 1989;107:1799-1803. 33. Marshall J, Trokel SL, Rothery S, Krueger RR. Longterm healing of the central cornea after photorefractive keratectomy using an excimer laser. Ophthalmology 1988;95: 141121. 34. Fantes FE, Hanna KD, Waring GO, Ill, et al. Wound healing after excimer laser keratomileusis (photorefractive keratectomy) in monkeys. Arch Ophthalmol 1990;108:665-75. 35. Malley DS, Steinert RF, Puliafito CA, Dobi ET. Immunofluorescence study of corneal wound healing after excimer laser anterior keratectomy in the monkey eye. Arch Ophthalmol 1990; 108:1316-22. 36. SundarRaj N, Geiss MJ III, Fantes F, et al. Healing of excimer laser ablated monkey corneas: an immunohistochemical evaluation. Arch Ophthalmol 1990; I 08: 160410. 37. Lohmann C, Gartry D, Kerr Muir MK, eta!. 'Haze' in photorefractive keratectomy: its origins and consequences. Laser Light Ophthalmol 1991 ;4: 15-34. 38. Talamo JH, Gollamudi S, Green WR, et al. Modulation of corneal wound healing after excimer laser keratomileusis using topical mitomycin C and steroids. Arch Ophthalmol 1991;109:1141-6. 39. Becker B, Mills DW. Corticosteroids and intraocular pressure. Arch Ophthalmol 1963;70:500-7. 40. Thygeson P. The unfavourable role of corticosteroids in herpetic keratitis. In: Brockhurst RJ, BoruchoffSA, Hutchinson BT, Lessell S, eds. Controversy in Ophthalmology. Philadelphia: WB Saunders, 1977;450-69. 41. Becker B. Cataracts and topical corticosteroids [editorial]. Am J Ophthalmol 1964;58:872-3. 42. Yablonski ME, Burde RM, Kolker AE, Becker B [editorial]. Cataracts induced by topical dexamethasone in diabetics. Arch Ophthalmol 1978;96:474-6. 43. Fitzsimmons TD, Fagerholm P, Harfstrand A, Schenholm M. Hyaluronic acid in the rabbit cornea after excimer laser superficial keratectomy. Invest Ophthalmol Vis Sci 1992;33: 3011-16. 44. Marshall J, Trokel S, Rothery S, Krueger RR. Photoablative reprofiling of the cornea using an excimer laser: photorefractive keratectomy. Lasers Ophthalmol 1986; I :21-48. 45. Tervo T, Tervo K, van Setten G-B, et al. Plasminogen activator and its inhibitor in the experimental corneal wound. Exp Eye Res 1989;48:445-9. 46. Lohmann CP, Gartry DS, Kerr Muir M, et al. Corneal haze after excimer laser refractive surgery: objective measurements and functional implications. Eur J Ophthalmol 1991;1: 173-80.