Contact Lens–Related Adverse Events and the Silicone Hydrogel Lenses and Daily Wear Care System Used

Objective  To investigate the incidence of adverse events related to the use of varying silicone hydrogel contact lens and lens solution combinations.

Methods  Individuals with myopia (N = 558) participated in 1 or more of approximately 40-participant trials in a matrix of 20 silicone hydrogel contact lens and lens-solution combinations. Visits were at baseline, 2 weeks, 1 month, and 3 months. The mean study completion rate was 90% of the expected participant-months (final data set: 840 lens-solution combinations and 2271 participant-months). Adverse events were reported as the first occurrence of each type per 100 participant-months for each lens-solution combination.

Results  The rate of all corneal infiltrative events (CIEs) was 3.1 per 100 participant-months (range, 0-10.5), and the rate of symptomatic CIEs was 1.7 per 100 participant-months (range, 0-10.5), including 1 case of microbial keratitis (0.04 per 100 participant-months). Rates for CIEs differed substantially among solution groups, with hydrogen peroxide having the lowest rate (0.6 per 100 participant-months; range, 0-0.9). The rate was 0.8 per 100 participant-months (range, 0-8.0) for superior epithelial arcuate lesions, which varied by lens type, 0.04 per 100 participant-months (1 case only) for corneal erosion, and 0.4 per 100 participant-months (range, 0-2.0) for contact lens papillary conjunctivitis, which was modified by type of solution. The rate of solution-induced corneal staining for all lens-solution combinations was 4.7 per 100 participant-months (range, 0-23) and varied significantly based on lens-solution combination (P < .001).

Conclusions  The frequency of adverse events varied with silicone hydrogel contact lens and lens solution combinations, with hydrogen peroxide having the lowest incidence of CIEs and solution-induced corneal staining, indicating that lens material and design, type of solution, and solution-lens interactions are likely contributing factors in this mode of lens wear.

Extensive investigation has been undertaken to determine the effects of contact lenses on the cornea and anterior segment. Owing to the risk of potentially sight-threatening sequelae, focus has centered on corneal infection, particularly with extended-wear lenses, which are associated with a higher risk of microbial keratitis (MK).^1,2 Daily wear, however, remains the major mode of use for conventional hydrogel materials and high oxygen–permeable silicone hydrogel (SiH) materials, originally introduced for extended wear.³

Although corneal infiltrative events (CIEs) other than MK are not sight threatening, they can result in discomfort and inconvenience that may cause discontinuation of lens wear. Silicone hydrogel lenses have also been associated with superior epithelial arcuate lesions (SEALs),^4-7 contact lens–induced papillary conjunctivitis (CLPC),^8,9 and other mechanical complications.^6,7,10

Recently, Jones et al¹¹ raised awareness about the issue of corneal staining resulting from the use of SiH lenses in combination with various solutions, and Andrasko and Ryen^12,13 attempted to predict such staining effects with a 2-hour provocative test. This focus has been intensified because of evidence suggesting that solution-induced corneal staining (SICS) is associated with corneal inflammation¹⁴ and decreased comfort.^13,15

In the absence of information regarding the frequency with which such occurrences might be expected, the present study aimed to establish incidence rates for adverse events (AEs) resulting from daily wear of SiH lenses combined with various lens-care products.

Data were collected from a series of ongoing, nonrandomized, open-label–dispensing clinical studies conducted at the Institute for Eye Research. All studies followed the same protocol, varying only by lens, solution type, and how often lenses were replaced. Commercially available SiH contact lenses and solutions were used daily for a 3-month period by 558 participants. Inclusion and exclusion criteria are listed in Table 1.

Details of lens types and solutions used during the studies are given in Table 2 and Table 3, respectively. Lens replacement schedules and the number of participants in each study are shown in Table 4. Approximately 40 participants were enrolled into each study, except for the study of lotrafilcon B lenses with Clear Care (CIBA VISION, Duluth, Georgia), which had 81 participants. Care was taken to record the number of neophyte (15.2%) and experienced (84.8%) wearers and to note the prestudy solution used, so as not to bias the outcome of forthcoming exposure. Lens wear was not interrupted before enrollment. Participants were permitted to enroll in more than 1 study, as long as 2 weeks had elapsed between trials. During this time, participants continued lens wear with their habitual lens and solution, or a more suitable combination prescribed by the optometrist. All procedures were conducted in accordance with the 2000 Declaration of Helsinki and were approved by the local area health ethics committee. Participants were advised of potential adverse reactions and completed written informed consent before enrollment.

All solution regimens consisted of a 5-second rinse (no rub) step before overnight disinfection; lenses were then taken directly from the case and inserted into the eye. Flat polypropylene cases were used in all studies using multipurpose solution. Examinations were conducted at baseline, 2-week, 1-month, and 3-month scheduled visits. Participants with persistent discomfort and AEs were seen ad hoc as appropriate. At each visit, compliance with wearing schedule and solution regimen was ascertained through self-report, and the anterior eye was examined using a Zeiss SL-120 biomicroscope (Carl Zeiss Meditech, Jena, Germany). All measures were collected by optometrists who underwent ongoing training to ensure consistency. Any CIEs, SEALs, CLPC, and corneal erosions were categorized according to the scheme proposed by the Institute for Eye Research^16,17 (Table 5).

Corneal staining using fluorescein sodium (Fluorets opthalmic strips, 1 mg; Chauvin Pharmaceuticals, Essex, England) moistened with 1 drop of normal saline was assessed using a cobalt blue light filter and a #12 yellow Wratton filter (Bausch & Lomb, Rochester, New York). The extent of corneal staining was graded according to a modified Institute for Eye Research scale (0, none; 1, <5%; 2, 6%-15%; 3, 16%-30%; 4, >30%) in 5 zones as previously described.¹⁸ Presence of SICS was recorded when diffuse punctuate staining (grade 1 and above) in at least 4 of 5 zones was noted.

Adverse events were defined as the first episode within each type of event for a participant at a follow-up visit (scheduled or ad hoc), apart from SICS, for which only scheduled visits were considered to avoid bias in studies with high numbers of unscheduled visits. Each participant was considered unique for each lens and solution combination. Once participants experienced an event, development of a different AE, but not a recurrence of the initial event, was included in the analysis. Incidence rates were calculated as the total number of first AEs within the study interval divided by the number of participant-months for that lens and solution combination. Rates are reported as the number of first AEs per 100 participant-months with the 95% confidence interval (CI) of the overall rate for daily-wear SiH contact lenses, which was calculated assuming a normal distribution. Where this could not be verified, the Poisson distribution was used instead.

The risk of a first AE was compared between lens and solution groups using logistic regression with a robust covariance estimator after accounting for participant repeats. If the main effect was significant using a likelihood ratio test, then pairwise comparison between the individual levels using Bonferroni correction was performed. Lens-solution combinations higher or lower than the 95% CI for the overall study rate were reported if interaction of the main effects was significant. Pairwise comparison was not performed because of the lack of statistical power.

Before logistic regression, the data were analyzed for possible confounding effects such as age, sex, smoking status, ethnicity, follow-up duration, prior lens and solution type used, and prior study participation. Significance was tested using a χ² test and a linear mixed model. Factors that were significantly different were considered as possible confounders (level 1 covariates). These covariates were factored into the multivariate analysis to adjust for possible confounding effects on the estimates of the main factors (lens or solution group). Extra variables such as replacement schedule, year, and season could not be tested as possible confounders because there was little or no variability within study cells. However, because these variables could have a potential effect, they were investigated on an exploratory level after adjusting for lens and solution type using a likelihood ratio test (level 2 covariates).

P ≤ .05 was considered statistically significant. Analysis was performed using SPSS statistical software, version 15.0 (SPSS Inc, Chicago, Illinois), and Stata statistical software, version 10.0 (Stata Corp, College Station, Texas).

The total sample comprised 558 participants enrolled in 1 or more trials in a matrix of 20 SiH contact lens and lens-solution combinations, resulting in 840 combinations of participant, lens, and solution. The distribution of participation in previous studies was positively skewed (P < .001); 21.0% previously participated in 1 study, 8.1% in 2, and 3.9% in 3 or more studies. The number of participants for each lens and solution type is shown in Table 4. Total daily wear exposure during the study was 2271 participant-months, a mean 90% (minimum 82%) of that expected for each lens and solution combination.

The demographic distribution is detailed in Table 6. The distributions of age, smoking status, and ethnicity were significantly different between the studies. Ethnicity and smoking status were not considered confounding factors because the distribution of ethnicity was not different between the studies (P = .16) after participants of “other” (ie, not white or Asian) race/ethnicity were excluded; when 1 study (lotrafilcon A with OPTI-FREE RepleniSH [Alcon Laboratories, Hünenberg, Switzerland]) in which no smokers participated was removed from the analysis, there was no difference in the proportion of smokers among the studies (P = .13). Age (factored as age ≤20 years), prior lens and solution group (SiH with hydrogen peroxide, SiH with multipurpose solution, non-SiH with hydrogen peroxide, non-SiH with multipurpose solution, or neophyte), and participation in previous studies were included in the multivariate analysis as possible confounders.

Participants experienced 70 first CIEs (all types) and 39 symptomatic CIEs. There were 18 SEALs, 1 corneal erosion, and 10 CLPC and 107 SICS events. The types of CIEs are summarized in the Figure. Data from the Institute for Eye Research Matrix Study are presented in Tables 7, 8, and 9.

Incidences of total and symptomatic CIEs are shown in Table 7. Total CIEs included MK, contact lens acute red eye, contact lens–associated peripheral ulcer, infiltrative keratitis, asymptomatic infiltrative keratitis, and asymptomatic infiltrates, whereas symptomatic CIEs included MK, contact lens acute red eye, contact lens–associated peripheral ulcer, and infiltrative keratitis only.

The incidence of total CIE was not significantly different between lens type (P  = .21) but was between solution types (P < .001). The effect of solutions did not interact with the lens type (P  = .18). Post hoc analysis indicated the incidence of total CIE with OPTI-FREE RepleniSH was higher than Clear Care (P < .001) and OPTI-FREE Express (P  = .01), and AQuify was higher than Clear Care (P < .001).

For symptomatic CIE, incidence was significantly different between lens types (P  = .01) and between solution types (P < .001). The effect of solutions did not interact with lens type (P = .30). Post hoc analysis indicated the incidence of syptomatic CIE with OPTI-FREE RepleniSH was higher than Clear Care (P = .01) and OPTI-FREE Express (P = .01), and lotrafilcon A was higher than lotrafilcon B (P = .03). Possible confounders were not significantly associated with total or symptomatic CIE (P > .05).

One case of MK, 1 case of contact lens–associated peripheral ulcer, and 37 cases of infiltrative keratitis occurred. Seven cases that occurred during the study resulted in referral to a general practitioner or ophthalmologist for treatment: 1 of bacterial keratoconjunctivitis and 5 of long-standing infiltrates, possibly viral or hypersensitivity in origin, in addition to the case of MK. The MK event resulted in loss of less than 1 line of visual acuity.²¹ Repeated events occurred for 14 cases (20.0%) of total CIEs and 6 (15.4%) of symptomatic CIEs.

A single event of corneal erosion occurred (senofilcon A [Johnson & Johnson, Jacksonville, Florida] with OPTI-FREE Express [Alcon] group).

The incidence of SEALs and CLPC are shown in Table 8. There was a significant difference for the incidence of SEAL between lens types (P < .001), but there was not between solution types (P = .08). The effect of lens type significantly interacted with solution type (P < .001). Balafilcon A used with AQuify and Clear Care was above the 95% CI of the overall rate, as was lotrafilcon A with OPTI-FREE Express and OPTI-FREE RepleniSH.

The incidence of CLPC was not different between lens types (P = .53) but was between solution types (P = .02). Solution type had a significant interaction with lens type (P = .002). The incidence of CLPC with Clear Care solution was above the 95% CI for 3 of the 5 lens types tested (galyfilcon A, senofilcon A, and balafilcon A) while lotrafilcon A was below the 95% CI. OPTI-FREE RepleniSH exhibited 2 (galyfilcon A and balafilcon A) and OPTI-FREE Express 1 (lotrafilcon A) lens-solution combinations above the 95% CI.

Previous study participation was not significantly associated with SEALs or CLPC (P > .05). Previous lens and solution group was associated with SEALs (P = .04) but not CLPC (P = .45). Age group was not associated with SEALs, but the effect of age could not be computed for CLPC since there were no cases of CLPC in wearers aged 20 years and younger.

Replacement schedule, chronological year, and season effect could not be computed for SEALs and CLPC alone since there were not enough cases of each event. The CLPC and SEALs were, therefore, combined into 1 “mechanical” category and analyzed. Replacement schedule was not associated with mechanical events (P  = .70). With all lens types combined, while the effect of season was not significant (P  = .07), chronological year was (P  = .002). When balafilcon A was removed from the analysis, chronological year was not significantly associated with mechanical events (P  = .90).

Only 16.5% of all inflammatory and mechanical events were diagnosed at unscheduled visits, which is likely owing to the asymptomatic nature of most events and the frequent visit schedule.

Table 9 details the incidence of SICS. The incidence of SICS was significantly different between lens types (P < .001) and solution types (P < .001), and there was a significant interaction between these factors (P < .01). Combinations using Clear Care or galyfilcon A were all below the 95% CI. Conversely, use of OPTI-FREE Express or OPTI-FREE RepleniSH was above the 95% CI with all lens types except galyfilcon A. The incidence of SICS used with AQuify solution was below the 95% CI for all lenses except balafilcon A, where the incidence was particularly high (23 per 100 participant-months). Age group, prior lens and solution use, and previous study participation were not significantly associated with SICS (P > .05). In addition, replacement schedule was not associated with SICS (P = .05). With all lens types combined, chronological year and season were associated with SICS (P = .03 and P = .001, respectively), however, when balafilcon A was removed from the analysis, there was no significance (P = .85 and P = .16, respectively).

This study is the first of its kind, to our knowledge, to report incidence of AEs for 20 varying daily wear SiH lens and lens solution combinations. With close to 90% of the lens-wearing population following a daily-wear schedule,³ even a low incidence of AEs is important. The sample of wearers enrolled in this study can be considered representative of the population of contact lens wearers; prescribing surveys have shown a similar age and sex distribution.³ Moreover, with the large sample size and high completion rates in this study, the reported incidence rates and 95% CIs can be considered to be reasonably robust.

Study results are consistent with previous reports associating specific lens and solution combinations with an increased risk of certain AEs.^11,12,20 Variation in lens solution AE rates suggests solution formulation differentially affects lens-eye interaction, an effect that is further modified by lens type. Minor CIEs can occur among people who do not wear contact lenses.^22,23 Although this is evidence that corneal inflammation can occur independent of lens wear, this study clearly demonstrates that lens type and solution influence the incidence of AEs among users of daily wear SiH lenses. Although many of the AEs in this study were asymptomatic, they are unlikely to be artifacts of observation because differences in incidence among lens solution combinations occurred. Although the etiology of asymptomatic CIEs are unknown, they are seen in non–lens wearers and probably represent corneal response to the local environment,¹⁰ which may be increased by lens type and/or lens and solution combination.

Few studies have documented CIE rates during daily wear of SiH lenses.^24-26 In a study by Morgan et al,²⁴ the annual incidence of severe keratitis was stated to be 0 (95% CI, 0-210.1) and of nonsevere keratitis it was 55.9 (9.9-309.7) events per 10 000 contact lens wearers per year. The study documented cases treated at a hospital, meaning that even nonsevere keratitis was likely to be symptomatic. In our study, considering symptomatic cases alone and converting to a 12-month period, the rate is considerably higher than that reported by Morgan et al as nonsevere keratitis (19.6 compared with 0.56 per 100 participant-years). A greater examination frequency and the absence of unsuccessful wearers discontinuing lens wear, together with a broader spectrum of symptom severity, is likely to have contributed to this difference in observed incidence. In another study, an Asian population of 87 previously successful lens wearers, with 6 months’ minimum experience, were given lotrafilcon A lenses to be used daily with habitual lens care for 1 month; no cases of infiltrate were reported.²⁵ The prior success of these wearers and the continuation of habitual lens care, as well as practitioners with less training in observing adverse events from several different sites, may have accounted for this apparently low rate. Two other short-term daily wear studies of experienced wearers that show decreases in infiltrate scores from baseline when refitting with SiH lenses^27,28 may also have been confined by these limitations. These studies contrast with the present study in which participants with active inflammation were not enrolled. In an analysis of preapproval and postapproval studies, Chalmers²⁸ highlighted how differences in apparent event rates may vary with study design.

Although pragmatic studies of this nature may be limited in that they are not masked or randomized, they reflect clinical practice and serve as an important marker of real-world experience.²⁹ Although the expansion of the SiH lens market when these studies were conducted meant a randomized study design was not possible, extensive statistical modeling has been used during analysis to account for confounding factors that may have been introduced as a result. There is a precedent for open-label studies in the assessment of contact lenses in clinical trials²⁵ because practitioners and participants often have experience with the lenses being tested and masking is not possible.

In contrast to the more usual participant-years, participant-months are reported because annualizing 3-month data does not account for dropout owing to unsuccessful lens wear. In a longer study with more successful wearers, this is likely to lower the incidence rate.

A higher proportion of symptomatic CIEs occurred with lotrafilcon A compared with lotrafilcon B, particularly with OPTI-FREE RepleniSH. Differences in lens design and modulus may have contributed to this disparity.

One case of MK was encountered that required urgent, aggressive pharmacological treatment. This case occurred in only 189 participant-years; however, no statistical inference can be drawn regarding the rate of MK among daily-wear lens users from this one case. The vast majority of events were not severe, and intervention for most CIEs was limited to temporary cessation of wear accompanied by normal saline eye rinses.

The use of OPTI-FREE RepleniSH resulted in a higher incidence of all CIEs compared with OPTI-FREE Express and Clear Care. Although they have many common ingredients (Table 3), OPTI-FREE Express and OPTI-FREE RepleniSH differ in terms of the concentrations of these reagents and TearGlyde (Alcon Laboratories), a lens-wetting agent, has also been added to OPTI-FREE RepleniSH. Differences in the multipurpose solution formulation seem to affect the level of irritation of the cornea; these effects become clinically apparent with daily wear.

Microorganisms are associated with inflammatory AEs among contact lens wearers.³⁰ In humans, no difference has been found in binding of Pseudomonas aeruginosa to different daily-wear and extended-wear SiH lens types with use of nonpreserved care systems.³¹ This contrasts with previous studies using preserved lens care solutions that showed an increase in binding of P aeruginosa to daily-wear SiH and hydrogel lenses.³² It is possible that microbial load or toxins at the epithelium may upregulate inflammation in SiH lens wear with preserved multipurpose solution.

Microblock (silver-impregnated) cases, supplied with AQuify solution (CIBA VISION), were not used in the trials to avoid possible bias of the solution effectiveness. Silver-impregnated cases have been shown to decrease lens case contamination in vitro, with moderate levels of microbial contamination, and in a clinical setting when used as directed for a 1-month period.³³ However, more than 60% of students in a 2007 Hong Kong–based study self-reported that they did not change their cases every 3 months.³⁴ Because the potential for contact lens wearers to use different cases than those supplied with solutions cannot be controlled for in the population, it is important to evaluate the solution as a stand-alone entity, as in the present study.

Ten cases of CLPC occurred in the studies. Although the number of cases of CLPC is too low to make any definite conclusions, the suggestion of greater incidence of CLPC with Clear Care and OPTI-FREE RepleniSH may indicate some potential of these solutions to render the lens more susceptible to greater mechanical or immunological stimulus for the development of CLPC.

The highest incidence of SEALs was observed among participants using balafilcon A and lotrafilcon A lenses. A lower incidence was seen among participants wearing galyfilcon A lenses, whereas no cases were reported with the use of lotrafilcon B or senofilcon A lenses. The fact that the modulus of lotrafilcon B is similar to balafilcon A (Table 2), but higher than galyfilcon A, suggests that the etiology of SEALs is not simply related to the modulus alone but involves other lens features. For example, the larger diameter and design of lotrafilcon B may result in a more even distribution of pressure onto the peripheral cornea, rendering it less likely to cause SEALs.

Previous participation was a significant confounder in SEALs, probably because no neophytes were affected. It could be inferred that repeated exposure of the cornea to lenses and/or solutions may increase the chance of developing SEALs. However, only 8 neophytes participated in the studies using balafilcon A, a lens type that showed a high proportion of SEALs. Had a higher number of neophytes worn balafilcon A lenses, it is likely that some would have developed SEALs, reducing the significance of this finding.

It is important to note that the incidence of SICS varied with lens and solution combination, as previously described.^12,13,20 One solution will not perform the same across all lens types, and a particular lens type will not give equivalent results with all solutions. Reference to the Institute for Eye Research Matrix Grid²⁰ is useful when changing lens and/or solution type to minimize the likelihood of SICS.

Benchmarking the performance during a 3-month period reveals differences in the incidence rates of AEs between various SiH lens-solution combinations. The occurrence of SEALs was influenced by lens type; total CIEs and CLPC varied by solution type; symptomatic CIEs were independently modified by both lens and solution; and SICS varied with the combination of lens and solution. Hydrogen peroxide exhibited the lowest incidence of CIEs and SICS. These results suggest that lens material and design, solution components, and solution-lens interaction play a role in the development of AEs among users of daily wear SiH lenses. In aiming to eliminate AEs with the use of SiH daily-wear lenses, further understanding of the interaction of the lens and solution and their effect on the ocular surface is crucial.

Correspondence: Nicole A. Carnt, BOptom, Institute for Eye Research, Level 3, Rupert Myers Building, University of New South Wales, Sydney, NSW 2052, Australia (n.carnt@ier.org.au).

Submitted for Publication: March 18, 2009; final revision received June 25, 2009; accepted June 29, 2009.

Funding/Support: This study was supported by CIBA VISION and the Institute for Eye Research.

Role of the Sponsors: A draft of the manuscript was provided to CIBA VISION; the only changes made as a result were grammatical in nature.

Additional Contributions: Isabelle Jalbert, PhD, and Lisa Keay, PhD, provided clinical direction for the studies from 2004 through 2006, Daniel Tilia, MOptom, was the principal investigator for some studies, Judith Flanagan, PhD, provided editing and writing assistance, and Klaus Ehrmann, PhD, and the Institute for Eye Research Technology Group provided the modulus measurements. The authors would like to thank the employees of the Institute for Eye Research Clinic from 2004 through 2008.