Corneal Haze Secondary to Corneal Collagen Cross-linking in Keratoconus Patients: Treatment and Outcomes

Abstract

PURPOSE:

The purpose of this study was to evaluate the outcome of patients presenting with mild-to-moderate corneal haze after undergoing corneal collagen cross-linking (CXL) for keratoconus (KCN) and their response to a proposed standardized topical steroid-based treatment.

METHODS:

This study included 19 eyes of 14 patients presenting with corneal haze after undergoing CXL for KCN. Corrected distance visual acuity, corneal thickness and Kmax values by Pentacam^® Scheimpflug tomography, as well as subjective corneal haze changes were evaluated before and after a topical steroid- and cyclosporine-based treatment.

RESULTS:

Visual acuity improved after the completion of the treatment by 0.043 logMAR (P = 0.017) and Kmax values decreased by 1.17D (P = 0.0024), while the corneal thinnest pachymetry remained stable. Data collected from the examiner’s slit-lamp examination description revealed that seven eyes had a decrease in haze compared to 12 eyes with stable or no changes in the haze.

CONCLUSION:

Our findings suggest an improvement in visual acuity and possible corneal flattening with decreasing Kmax after the completion of topical steroids with taper course treatment in patients suffering from corneal haze post-CXL. This paper also highlights the importance of postcross-linking anti-inflammatory treatment and close follow-up.

Introduction

Keratoconus (KCN) is a noninflammatory degenerative ectatic condition that affects the eye’s cornea. It is associated with characteristic features like stromal thinning resulting in the conical transformation of the cornea, irregular astigmatism, and impaired vision.[1] The condition’s incidence varies according to the geographic location, with estimates in the general population being approximately 1/2000. There is a clinical impression backed by some preliminary data that in Lebanon, where this study is conducted, KCN is more frequent with a reported prevalence of 3.3%.[2] Early stages of the disease typically start in the adolescent years and progress throughout the second and third decades of life until gradual stability is reached.[3] The mainstay of early treatment focuses on the correction of the ectasia-induced refractive error by rigid contact lenses and may be followed by corneal ring insertion or corneal transplantation.[4] The only treatment aimed at slowing or halting the progression of the disease is ultraviolet (UV)-A light associated with riboflavin collagen cross-linking (CXL).[1] Like any procedure, CXL is associated with early- and late-onset complications such as pain, keratitis, corneal opacity, corneal edema, endothelial damage, persistent epithelial defect, and corneal melt.[5] The purpose of this retrospective descriptive case series was to evaluate the outcome of patients presenting with mild-to-moderate corneal haze post-CXL and their response to a proposed standardized topical steroid-based treatment.

Methods

Data in this study were collected and analyzed by retrospective review of medical records of 14 patients (total of 19 eyes) presenting to our referral center for 3 years with corneal haze post-CXL for KCN. The surgical technique and patient preoperative objective data are unavailable since all CXL procedures were performed elsewhere. Patients were referred to our center after an average of 4.46 years (standard deviation (STD)= 2.75) since their CXL procedure. The main inclusion criteria were: any patient with mild-to-moderate central corneal haze (Grade 1–3 on slit-lamp examination) after the CXL procedure associated with any type of visual disturbance. Patients with interfering pathologies (such as keratitis and corneal grafts) or higher grades of haze were excluded.

All patients were offered 2-month cycles of the following standardized topical treatment:

• Loteprednol etabonate ophthalmic suspension 0.5% starting at four times/day for the first 2 weeks, and then tapered to three times/day for 2 weeks, then two times/day for 2 weeks, and finally once daily for two more weeks

• Cyclosporine ophthalmic suspension 0.05% twice per day for 2 months

• Preservative-free sodium hyaluronate-based artificial tear substitutes four times/day for 2 months

• Carbomer 0.2% eye gel at night for 2 months.

The following variables were analyzed before and after the proposed 2-month treatment:

• Corrected distance visual acuity (CDVA in logMAR) measured using a Snellen chart

• Slit-lamp biomicroscopic subjective description of haze grade assessed by the same examiner

• Corneal steepest Corneal steepest keratometry readings (Kmax) pre- and posttreatment from corneal tomography (Pentacam^®).

Results

A total of 14 patients (19 eyes) with a mean age of 33.7 years ± 7.8 years (range: 20–44 years) were included in this descriptive case series. Pertinent patient data are summarized in Table 1.

Table 1.

Clinical data of patients presenting with corneal haze postcross-linking before and after the proposed treatment cycle

Case number	Age (years)	Years since CXL	Scarring (OD, OS, OU)	Before treatment			Posttreatment
				CDVA (logMAR)	Kmax (D)	Corneal Thickness (µm)	CDVA (logMAR)	Kmax (D)	SLE changes in haze
1	24	5	OU	0.7	56.9	353	0.7	55.1	Decrease
		5		0.87	52.3	301	0.7	50.4	Decrease
2	35	10	OS	0.52	48.67	482	0.52	47.65	Stable
3	32	4	OU	0.05	47.7	425	0.05	45.44	Stable
		4		0.15	43.8	439	0.15	41.8	Stable
4	39	4	OS	1.3	50.42	395	1.3	49.7	Stable
5	36	0.3	OD	0.4	52.34	485	0	49.34	Decrease
6	32	2	OS	0	50.25	413	0	52.25	Stable
7	27	1	OD	0.3	53.64	421	0.3	51.8	Decrease
8	34	6	OS	0.18	49.95	391	0.18	48.9	Stable
9	42	1	OU	0	53.82	438	0	50.6	Decrease
		0.5		0.3	54.47	424	0.18	51.2	Decrease
10	20	6	OU	0.3	46.73	466	0.3	48.1	Stable
		6		0.4	46.54	452	0.4	46.42	Stable
11	42	9	OD	0.3	50.25	344	0.3	49.5	Stable
12	44	4	OU	0.4	50.02	424	0.4	49.9	Stable
		4		0.15	49.67	467	0.15	49.67	Stable
13	24	5	OS	0	53.3	517	0	51	Decrease
14	40	8	OD	0.4	50.36	330	0.4	50.06	Stable

CXL: Collagen cross-linking, CDVA: Corrected distance visual acuity, SLE: Slit lamp exam, logMAR: Logarithm of the minimum angle of resolution

A comparative table of the pre- and posttreatment variables can be seen in Table 2.

Table 2.

Comparative table of means and standard deviations before and after our proposed steroid-based treatment

Parameter	Mean±SD		P
	Pretreatment (n=19)	Posttreatment (n=19)
CDVA (logMAR)	0.3537±0.32578	0.3105±0.32581	0.017
Kmax (D)	50.59±3.1559	49.41±2.8166	0.0024

CDVA: Corrected distance visual acuity, SD: Standard deviation

Data collected from the examiner’s slit-lamp examination description revealed that seven eyes have a decrease in haze compared to 12 eyes with stable or no changes in the haze.

There was a statistically significant 0.043 logMAR improvement in CDVA (P = 0.017) and 1.17D reduction in Kmax (P = 0.0024) when comparing pre- and posttreatment values as reported in Table 2. The corneal thinnest pachymetry was stable before and after treatment protocol with no significant change.

Discussion

Our main goal in this study was to observe the outcome of CDVA, corneal Kmax readings, and anatomical haze changes after a topical steroid-based 8-week cycle treatment was delivered.

While cross-linking offers proper halting of KCN, associated complications like corneal haze can have large effects on visual acuity. It is postulated that stromal scar or corneal haze located away from the visual axis has more favorable outcomes.[6] Our study showed that in most eyes, CDVA improved or stayed the same, which was reflected by a mean of 0.0432 logMAR reduction. To note, all eyes had mild-to-moderate central corneal haze formation. We believe that the improvement in CDVA is a result of both reduction of Kmax with subsequent corneal flattening as well as an anatomical decrease in the haze. CXL promotes stromal fibrosis with resultant corneal flattening and Kmax decrease.[7] Long-term cross-linking follow-up has also shown that his process may extend for years.[8] Our study showed an average decrease of 1.18 D, which may be explained by the decrease in haze as well as the long-term corneal flattening effect of CXL.

Anterior stromal opacity may appear in the 1^st weeks to months after CXL in 10%–90% of patients extending up to 60% stromal depth.[9] The descriptive changes assessed by the examiner in this study revealed an improvement and corneal haze decrease in seven eyes compared to stable and no significant change on slit-lamp examination in 12 eyes. Since quantification and grading of haze are subjective and observer dependent, we opted to present our finding as a description of the corneal haze progression either as increasing, stable, or decreasing at the end of the treatment protocol. Scheimpflug densitometry provides objective and accurate quantification of corneal haze but was not available for this study.

Eyes with thinner corneas and higher Kmax values at the apex of the KCN are more predisposed to the development of haze after CXL.[10] While no predictors of response could be identified, it is worth noting that eyes with reported haze decrease in response to the topical treatment tended to have higher Kmax values with no correlation with pachymetry.

It is established that corneal haze first appears at 1 month after CXL, plateaus between the 3^rd and 6^th month after which significant regression begins and extends until 1 year after the procedure.[11,12] Our study found that persistent corneal haze beyond a year after cross-linking can be reduced with topical steroids with steeper and thinner corneas showing greater response. A prospective study done in 2009 by Koller et al.[13] postulated that haze resulting from CXL extends deep into the corneal stroma compared to the superficial haze seen after photorefractive keratectomy (PRK) and may be associated with the depth of CXL and loss of keratocytes. This could also explain why the thinner cornea was to have a better response to a topical steroid that seems to have better penetration and greater reach to the posterior and deeper stromal haze.

Risk factors for haze formation following PRK have been reviewed in some studies and could play a similar role in predisposing CXL patients to this common complication. Some of the factors listed were well established in our study group such as astigmatism larger than 1.25 D, increased UV exposure in our moderate-to-high UV index region, and brown irides.[14] Other factors mentioned included ocular surface disorders, pregnancy, oral contraceptives, and previous corneal surgery.[14] Our study, nonetheless, lacked further preoperative and operative conditions and details making it difficult to correlate the aforementioned factors to haze formation in the reported patients.

Topical steroids have a direct inhibitory effect on stromal fibroblasts through decreased DNA synthesis and keratocyte activity, leading to reduced stromal healing and unwanted haze formation.[15] Our descriptive study revealed that an 8-week taper course of topical steroids can decrease stromal haze resulting from a CXL procedure. Furthermore, a recent study comparing the short-term and long-term steroid use for corneal haze after PRK showed no significant difference, therefore recommending a shorter course of 1 month.[16] In our cases, a rather moderate regimen of 8 weeks was used to seek any noticeable decrease in the haze with the addition of topical cyclosporin throughout the duration of treatment. We believe that the duration of treatment chosen agrees with the aforementioned study by implementing an ideal duration to maximize the benefits without exposing the patient to unnecessary long-term topical steroids.

Moreover, six of the 19 eyes were noted to have a corneal thickness below 400 μm, which is less than the recommended thickness for performing CXL safely.[17] Unfortunately, we had no data about the exact technique used in such cases, which could have led to haze formation with or without endothelial damage.

Conclusion

Our descriptive study suggests that completing a course of topical steroids with taper can decrease stromal haze resulting from cross-linking. This was shown to be effective even with latent haze presentations resulting in a better visual outcome for patients. This paper also indirectly highlights the importance of postcross-linking extended anti-inflammatory treatment and close follow-up. However, we strongly believe that a large prospective randomized clinical trial would validate our proposed treatment cycle and where our study falls short such as a larger study group, standardized and well-reported cross-linking technique, objective haze quantification, and repeated postsurgical follow-ups.

Financial support and sponsorship

This article was not funded by a third party in any form.

Conflicts of interest

There are no conflicts of interest.