Deciphering Corneal Staining Scales
A look at the impact of grading scales and study designs on corneal staining results in clinical studies.
Dr.Lasswell is Director, Clinical Research, Eye Care and Refractive Surgery, at Advanced Medical Optics, Inc. He has over 30 years of ophthalmic industry experience, has given more than 50 professional presentations and published several papers related to contact lenses and refractive surgery.
Mr.Huth is a Research Investigator for Eye Care R&D at AMO. He holds 14 US patents and has 13 patent applications pending.
Ms. Tran works in Eye Care R&D for AMO. She graduated from UC Irvine with a degree in Chemistry in 2005.
By Lynn Lasswell, OD, FAAO, Stan Huth, MA, & Denise Tran, BS
Corneal staining is a valuable clinical tool for assessing corneal epithelial integrity at the slit lamp. Recent reports of clinically significant corneal staining with silicone hydrogel lenses have highlighted the importance of understanding the interaction of lenses, lens care solutions and the corneal surface.
Understanding these relationships and integrating the study findings into the current body of knowledge requires comparable data from multiple studies. This is difficult to achieve when investigators use different patient populations, different scales to generate staining scores and different benchmarks for staining intensity and clinical severity.
Here we’ll present a summary of the major corneal staining grading scales, discuss a study that uses its own internal grading scale to evaluate a variety of multipurpose solution and lens combinations, provide previously unpublished data from six studies using two different corneal staining scales for one lens solution and several lenses, and report on other contributors to corneal staining.
We have no universal standard for grading the extent or severity of corneal staining. The major grading scales are not in congruence with one another. The scales currently employed are subject to differences in grading precision between different observers (inter-reader reliability) and between one test and another for the same observer (intra-reader reliability). In an interesting study that examined how investigators assess corneal staining, Begley et al (1996) reported that there was significant inter-reader variability in how the investigators within a multicenter study arrived at the final corneal staining score, even though all three sites were provided with the same guide for assessing corneal staining. The authors demonstrated that, depending upon the process used, two investigators could report significantly different scores for the same level of corneal staining.
In one study presented here, a multipurpose solution marketed under different names but used with the same silicone hydrogel lens and evaluated using the same scale, produced variations in staining results.
Practitioners should pay careful attention to the grading scales, study designs, populations studied and the consistency of results from several studies before relying on the conclusions presented in any given report.
Traditional Corneal Staining Grading Scales
Corneal staining grading scales were developed to help clinicians monitor changes of the cornea and choose a course of action. Clinicians observe the corneal surface under the slit lamp, compare it to written descriptions, photographic, artist-rendered or computer-generated images of different severities of corneal staining and assign a grade based on one or more elements of the staining and the number of zones of staining on the cornea. The grading systems help clinicians assess which corneal staining patterns are within the range of normal and which are pathological and in need of therapeutic intervention.
Each of the major grading systems employs a method of dividing the cornea into zones and for evaluating one or more staining variables in each zone.
In the five-zone model, Zone 1 is a circle in the center of the cornea and the four equal segments of the ring surrounding this central zone are Zones 2 through 5 (superior, temporal, nasal and inferior zones). The observer makes an estimate of the zonal area involvement and calculates the staining score based on the number and/or type of staining variables.
Efron The Efron system is based on the work of Nathan Efron, DSc, MCOptom, FAAO, FCLSA, FBCLA, FIACLE (currently based at the University of Queensland, Australia) when he was at the European Centre for Contact Lens Research, Department of Optometry and Neuroscience, University of Manchester in the United Kingdom.
The Efron system uses one variable, the degree of staining per zone based on an ordinal scale of 0 to 4: 0=no staining, 1=trace staining, 2=mild staining, 3= moderate staining, 4=severe staining. The Efron system has been validated for clinical use with an expected accuracy of ±1.2 grading scale units, a rather large range of error for a four-point scale.
CCLRU The scale developed by the Cornea and Contact Lens Research Unit (CCLRU), School of Optometry and Vision Science, The University of New South Wales, Sydney, Australia, calls for three variables per zone: type, depth and extent of surface area staining.
Figure 1 shows the CCLRU punctate staining types. Type 0 means there is no staining. Type 1 is micropunctate; Type 2, macropunctate; Type 3, coalescent macropunctate staining; and Type 4 is a coalescent patch of 1mm or greater in size.
The other two variables in the CCLRU system are depth and extent of staining. If there is no staining, the depth is graded as 0. Superficial epithelial involvement is Grade 1. The presence of a stromal glow within 30 seconds is Grade 2. An immediate localized stromal glow is Grade 3, and immediate diffuse stromal glow is Grade 4.
If there is no staining, the extent is graded as 0. From 1 to 15 percent of surface involvement is Grade 1, 16 to 30 percent surface involvement is Grade 2, 31 to 45 percent surface involvement is Grade 3 and 46 percent or greater surface involvement is Grade 4.
The CCLRU scale uses the zone of greatest staining to determine clinical significance. Under their criteria, corneal staining is clinically significant when it is persistent, its type is greater than Type 2 (macropunctate) and/or its depth is greater than Grade 1 (superficial epithelial involvement) and/or its extent is greater than Grade 1 (1 to 15 percent surface involvement) in a given zone. Micropunctate staining is considered not clinically significant by the CCLRU unless it involves more than 15 percent of the corneal surface.
Figure 1. Types of punctate fluorescein staining in CCLRU standards superimposed on a photo from the CCLR group. Image copyright is owned by the Centre for Contact Lens Research, School of Optometry, University of Waterloo, Ontario, Canada. Used with permission.
CCLR/Global Staining Score Lyndon Jones, PhD, FCOptom, FAAO, at the Centre for Contact Lens Research (CCLR), School of Optometry, University of Waterloo, Ontario, Canada, expanded on the CCLRU system to make it more sensitive by changing the ordinal scale of 0 to 4 to an integer scale of 0 to 100. Under this system, the type of staining in each zone is graded on a 0 (none) to 100 (total) scale. The mean outcome measure, the Global Staining Score, is the product of the type of staining in each of five zones (one central and four peripheral) times the percentage area of the zone with the staining. Under this system, the scale ranges from 0 (none per zone) to 50,000 (total staining in all 5 zones). Recently, the scoring was modified slightly in an attempt to normalize the scores to represent a typical score for a given sector. The Global Staining Score is now divided by 5, for a maximum average sector staining score of 10,000. For example, one corneal sector with micropunctate staining (score 25) over 10 percent of its surface would have a total score of 250 (25 × 10). If all have sectors had the same score, the total score would be 1,250 (250 × 5), but this would be normalized to 250. An average sector score of less than about 1,200 (or total Global Staining Score of 6,000) is considered clinically insignificant. Figure 2 shows the representative staining score for each of the CCLR types of punctate fluorescein staining.
Differences among the leading grading systems appear in Table 1.
Difficulty in Comparing Staining Scales in Studies
The difficulty in drawing conclusions from various studies arises when these studies utilize different staining scales and methodologies. This is illustrated in a study presented as a poster at the Association for Research in Vision and Ophthalmology (ARVO) meeting in May 2006 and at the American Optometric Association (AOA) meeting in June 2006 (Andrasko et al). Researchers used a modified CCLRU scale, the pairs of lenses and solutions have different numbers of subjects, and the same solution marketed under different brand names produced different staining scores.
This is a double-masked, randomized, crossover study at a single site with a planned enrollment of 200 patients in a series of one-week studies with different combinations of contact lenses and multipurpose solutions. Researchers evaluated fluorescein staining as well as lens and solution comfort at baseline (after 15 minutes of lens wear) and at two and four hours of lens wear. The investigators use their own corneal staining scoring system that differs from the aforementioned quantitative scales and also is not comparable to the CCLRU qualitative scales. The staining score is based on area of staining in each of the five corneal regions. The investigators assessed the percentage of each stained corneal region and averaged the percent area scores for the five regions to give a composite staining value. The average staining levels for the worse eye of each subject are reported. This is a measure of area stained, not a qualitative measure of type or depth of the staining.
The researchers planned the pilot study for nine to 14 patients in a two-period crossover study and the other studies for 30 patients each, one a two-period and the other a five-period crossover study.
The number of subjects and staining scores for several of the pairings were changed without notice or explanation between two recent updates of this study’s online staining grid. The sample size for the PureVision (Bausch & Lomb) lens/Opti-Free Express solution (Alcon) combination was 36 in the Oct. 17 update, or six more subjects than most of the other study combinations, and was inexplicably reduced to 30 at the Nov. 9 update. The staining score of 6 percent with 36 subjects was lowered to 4 percent with 30 subjects.
Also in this study, the same care solution formulation marketed under three different brand names returned different outcomes. Wal-Mart’s Equate brand and the Target brand of lens care solutions are the same formulation as ReNu MultiPlus multi-purpose solution (B&L). The formulation was associated with 24 percent (N=29, ReNu MultiPlus MPS), 28 percent (N=30, Target MPS) and 41 percent (N=30, Equate MPS) corneal staining with O2Optix lenses (CIBA Vision). When paired with Night & Day (CIBA Vision) lenses, corneal staining was 24 percent (N=30, ReNu MultiPlus MPS), 24 percent (N=30, Target MPS) and 36 percent (N=30, Equate MPS). Different patient populations or, as other studies have noted, discrepancies in inter- and intra-reader repeatability, can produce different results.
Figure 2. Global staining scores are shown for each type of punctate fluorescein staining (CCLR standards). Image copyright is owned by the Centre for Contact Lens Research, School of Optometry, University of Waterloo, Ontario, Canada. Used with permission.
Standard Staining Scales, Different Solutions
The above study reported poor results for Complete MoisturePlus Multi-Purpose Solution (Advanced Medical Optics, Inc.). We report six previously unpublished studies with Complete MoisturePlus solution and silicone hydrogel lenses (Table 2). All were based on recognized staining scales. Studies 1, 2 and 3 were based on the CCLR/Global Staining Score scale and Studies 4, 5 and 6 on the Efron scale. The Efron scores showed no significant staining for O2Optix, Acuvue Advance (Vistakon), Night & Day and Acuvue Oasys (Vistakon) contact lenses. Similarly, the more sensitive Global Staining Scores showed no significant corneal staining with O2Optix, Acuvue Advance and Night & Day contact lenses.
Supporting these AMO studies is a study from Bausch & Lomb. B&L found insignificant levels of corneal staining in a three-month, 25-site prospective study of corneal staining with silicone hydrogel lenses paired with ReNu MultiPlus (N=151), Opti-Free Express (N=150) or Complete MoisturePlus (N =156) multipurpose solutions. The researchers conducted slit lamp evaluations at two weeks, one month, two months and three months and at any unscheduled visit. Opti-Free Express solution had 92.7 percent with no corneal staining, 6.3 percent with Grade 1 (trace), 0.8 percent with Grade 2 (mild), 0.2 percent with Grade 3 (moderate) and 0 percent with Grade 4 (severe) staining. ReNu MultiPlus solution had 90.5 percent with no corneal staining, 8.8 percent Grade 1, 0.5 percent Grade 2 and 0.1 percent Grade 3. Complete MoisturePlus had 91.6 percent with no corneal staining, 8.2 percent Grade 1, 0.2 percent Grade 2 and 0 percent Grade 3.
An earlier study by Amos and colleagues (2004) showed solution-related differences in corneal staining in 25 contact lens wearers switched to Night & Day silicone hydrogel lenses who used ReNu MultiPlus solution for the lens in one eye and Focus Aqua (Aquify in the United States, [CIBA Vision]) solution for the other. Mean corneal staining was 0.8 at baseline for ReNu MultiPlus eyes and 0.7 in the Focus Aqua eyes (P=NS) on a scale of 0 to 4. Corneal staining was clinically and statistically significantly higher with ReNu MultiPlus than with Focus Aqua at both the two week (1.5 vs 0.8) and one month (1.2 vs 0.5) visits (P=0.0001 for both). The disinfecting agent and preservative for both these solutions is polyhexamethyl biguanide (PHMB). PHMB is present in similar amounts in the two solutions, suggesting that other constituents in the formulation play a role in the development of corneal staining.
Etiologies of Corneal Staining
Corneal staining is linked to many factors, not all of which are contact lens-related, and some of which may represent normal levels of staining in healthy eyes. Schwallie et al (2001) identified corneal staining in healthy eyes of non-contact lens wearers as 0.5 on a 0 to 4 scale with 0.5-step differences. Dundas et al (2001) studied 102 subjects who were non-contact lens wearers or who had no recent contact lens wear. They found some degree of fluorescein staining on 79 percent of subjects’ corneas based on the CCLRU grading scale. Half of the subjects showed staining in the inferior or superior zones and 5 percent had staining in the central zone.
Corneal staining also occurs in successful hydrogel lens wearers and may result from issues other than the care system, including the experimental conditions of the study. Begley and colleagues in Columbus, Ohio (1996) conducted a multicenter study with 98 full-time contact lens wearers using different lens care systems and soft contact lenses. Grading was on the 0 to 4 scale with half-step differences for each of five corneal zones. The average staining grade was 0.50 in both eyes, but the between-eye values were significantly different, 0.57 for the right and 0.44 for the left eye (P=0.001).
Nichols and colleagues (2002) examined 500 fulltime successful hydrogel contact lens wearers and found corneal staining in at least one eye in 55.7 percent of wearers. Eight percent had moderate to severe staining. The authors concluded that moderate to severe staining was associated with noncompliance with the care systems, a conventional rather than a planned or disposable lens replacement schedule, and lenses with powers greater than −3.00D.
Topically Applied Solutions
Fluorescein appears to increase corneal permeability with repeated use, as does the topical anesthetic proparacaine hydrochloride. Josephson and Caffery (1988) looked at five quadrants of the cornea for the presence or absence of corneal staining after sequential instillations of 9mg fluorescein and following administration of proparacaine hydrochloride 0.5% with benzalkonium chloride (BAK) in 148 eyes of 74 healthy subjects who did not wear contact lenses. A single exposure to 9mg of fluorescein resulted in 26 eyes (17.6 percent) with staining and 122 eyes (82.4 percent) with no staining at 60 seconds. The anesthetic was instilled followed two minutes later by fluorescein. Of the stained eyes, 10 (38.5 percent) remained at the same level and 16 eyes (61.5 percent) had increased staining. Of the unstained eyes, 59 (48.4 percent) continued to have no staining and 63 eyes (51.6 percent) developed staining.
Corneal staining may be a sign of corneal disease and can identify patients who are at risk of developing corneal infiltrates. Papas and colleagues (2006) at the Institute for Eye Research in Sydney, Australia, studied 609 subjects over 16 clinical trials using various lens-solution combinations for up to three months. One out of 10 subjects who had staining had an infiltrate, and subjects who had diffuse punctate staining in at least four of five areas of the cornea were three times more at risk of developing a corneal infiltrative event. Hydrogen peroxide solutions had low toxic staining rates.
The studies listed above are all short-term. Few studies have looked beyond the first three months of wear to examine persistent staining. Santodomingo-Rubido (2006) identified a significant subjective increase in corneal staining over the first three months of silicone hydrogel lens wear in 45 patients using Opti-Free Express MPS or ReNu MultiPlus MPS and followed for 18 months (p<0.05).
Differences in grading scales and study designs make it difficult to compare results across clinical studies. Investigators use different scales to generate corneal staining scores, and these scales are subject to differences in grading precision. A study comparing numerous contact lens and contact lens care solution combinations produced disparate corneal staining results for the same care solution marketed under three different names. This disparity highlights the need in evaluating clinical study reports to pay attention to the grading systems used, the populations studied, and the consistency of results from several studies. CLS
To obtain references for this article, please visit http://www.clspectrum.com/references.asp and click on document #137.