Meta-analyses and systematic reviews are critical to deciphering published and unpublished data common to the same topic. A recent comment in the Lancet stated that clinical trials done in the absence of such reviews and summary measures are scientifically and ethically improper.1 Since silicone hydrogel (SH) contact lenses were introduced for up to 30 days of continuous wear, they have been scrutinized for their safety. Thus, a summary measure of the risk of certain complications associated with SH lenses compared to their traditional low oxygen permeable (Dk) counterparts, is an important contribution to the literature.
The clinical performance of SH lenses clearly demonstrates an improvement compared to low Dk lenses on a number of physiological markers making them desirable for extended wear. Improvements in ocular health are noted in all corneal layers as well as in the conjunctival response to lens wear.2–13 In short, SH high Dk materials are a tremendously positive addition to the contact lens practitioner’s armamentarium as they have been shown to effectively eliminate the hypoxic impact of continuous wear on the corneal physiologic status.
The United States Food and Drug Administration (FDA) has mandated large scale post approval studies for 30 day continuous wear SH lenses to monitor them for safety. The first of these studies to be published14 documented an overall rate of presumed microbial keratitis for a 30-night SH lens similar to that reported for low Dk extended wear lenses worn for fewer consecutive nights.15,16 However, it is still questionable whether high Dk SH lenses are safer than their low Dk counterparts in regards to other forms of less severe keratitis. The question of safety remains as the clinical and research communities continue to report a variety of adverse responses with continuous wear of SH lenses. Risks of acute or chronic inflammation still remain because they are not caused by corneal hypoxia alone. Large scale epidemiological studies show that SH continuous wear has the same absolute risk for microbial keratitis as low Dk extended wear.14,17 One study found the overall incidence of contact lens-related keratitis which required hospital treatment was about the same for SH and low Dk extended wear,18 however, the clinical severity of the keratitis was lower in the SH group when compared with the low Dk group.19
In at least 17 peer-reviewed studies since 2000, the efficacy and safety associated with overnight use of SH lenses have been reported. With the exception of rare cases of MK, usually milder adverse responses such as mechanical and inflammatory events have been documented. The incidence estimates vary widely among these studies. For example, the range of corneal inflammatory events with SH lenses reported around the globe claim that the incidence proportions range from about 2% to over 21%.20–25 When compared with traditional hydrogels, SH lenses enable healthier corneas which may render greater resistance to inflammation or infection,9 and they show higher tear elimination rates which may enhance flushing of potentially antigenic postlens debris.26 Therefore, the risk of corneal inflammatory rates with SH lenses should theoretically be reduced compared to traditional hydrogels. A few studies have directly compared SH with low Dk disposable hydrogels.20,21,27–31 All but one of these studies utilized SH lenses for up to 30 days continuous wear and low Dk lenses on a 7-day wear schedule, and documented a higher rate of corneal inflammatory events in the SH group (although the differences compared to low Dk lenses were not statistically significant). The other study utilized both SH and low Dk lenses on a 7-day extended wear schedule, and interestingly, had an insignificant but slightly higher rate of inflammatory events in the low Dk lens group.28
Meta-analysis is a systematic approach to identifying, appraising, synthesizing, and (if appropriate) combining the results of relevant studies to arrive at conclusions about a body of research.32 This statistical method utilizes the results of a number of different studies to provide a larger sample size for evaluation, provide a consensus of findings in the medical and research communities, and assess the reproducibility of these findings. The purpose of this study was to perform a meta-analysis to combine the relevant studies to evaluate the difference in risks of corneal inflammatory events, excluding microbial keratitis, between SH and low Dk hydrogel extended lens wear.
The methods and reporting used in this meta-analysis follow the recommendations published in the Journal of the American Medical Association by the MOOSE Group.32
Search Strategy and Data Extraction
A systematic search of literature was conducted using the MEDLINE (PubMed), EMBASE, Cochrane, and Reference Sight databases. MEDLINE is the electronic version of Index Medicus, EMBASE is the Excerpta Medica database available on the Dialog search system, the Cochrane Library consists of a regularly updated collection of evidence-based medicine databases including The Cochrane Database of Systematic Reviews, and Reference Sight is an online collection of references and resources on contact lenses, refractive surgery, and external ocular disease developed and maintained by the Centre for Contact Lens Research at the University of Waterloo. Additional searches included abstracts from the annual meetings of the American Academy of Optometry as well as the Association of Research and Vision in Ophthalmology (ARVO), and the FDA websight (http://www.fda.gov/cdrh/) for SH lens approvals, and their corresponding Summary of Safety and Effectiveness Data. Additional studies were identified by engaging in personal communication with established scientists in the field and by reviewing references of retrieved articles.
All searches encompassed the years between 1990 and February 2006. Search terms included a primary search of the subject headings “silicone hydrogel contact lenses” or “extended wear contact lenses”. A second level search combined one of the first two subject headings with one of the following terms: corneal infiltrates, corneal inflammatory events, complications, or adverse events.
Data extraction included the sample size for each arm of the study (SH or low Dk disposable hydrogel), number of subjects, events, or eyes developing corneal inflammatory events in each arm, country of study, type of material(s) worn and whether prototype lenses were used, length of extended wear, frequency of disposal, solutions or care regimens used, daily wear adaptation period, study design, eligibility criteria, age, sex, race, and refractive error of subjects, percentage and types of adapted wearers, length of study, sponsor of study, definition used for corneal inflammatory events, and an assessment of whether single or multiple events were reported. Both authors independently extracted data following a preconstructed customized form, and any disagreement was resolved by discussion before completing fields in the database. In studies where information was not provided in the article, but was considered to be important in the analysis, individual authors were queried.
Prospective or retrospective cohort studies and randomized controlled trials in any language were eligible. We used a broad inclusion criteria in relation to heterogeneity of designs and variable definitions, and then performed analyses relating design or other study feature to outcome. Case–control and most prevalence studies were excluded, because adequate numerators and denominators for incidence rates could not be ascertained from the data.
Studies must have clearly stated the lens materials used (low Dk materials could have been classified by their FDA lens group). Studies including prototype lenses were not excluded, however, prototype and older design lenses were adjusted for in subsequent analyses. In studies published before 2003 using lotrafilcon A lenses, it was assumed an older design and single base curve was used. (Ciba Vision released a new design and two base curve options for lotrafilcon A in January of 2002, before that only a single 8.6 base curve was available). Studies must have included normal subjects with active follow-up of at least 4 months in each arm using an extended wear regimen. Studies where extended wear lenses were used for therapeutic use were excluded. Studies reporting on nondisposable low Dk hydrogels or ulcerative keratitis were excluded. Multiple articles from the same author or institution were closely analyzed to ensure the data extracted were not duplications or accumulations of data from other studies. The process of selecting the 23 studies that met our inclusion criteria is shown in Fig. 1.
All studies that reported on the number of corneal inflammatory events over a defined period of time in a defined subset of individuals were included regardless of the definition used for the inflammatory event. However, data were extracted from each study on the method of definition of a corneal inflammatory event, and a study was then classified into one of two strata. The purpose of stratifying by event definition was to provide distinct point estimates and risk ratios for each stratum because the incidence rates were expected to be much higher if the authors reported all possible inflammatory findings compared to reporting only a subset of inflammatory events.
The first stratum defined an event as any corneal inflammatory finding documented on a scheduled or unscheduled follow-up visit regardless of the slit-lamp grade.21,24,29,30,33–37 Corneal inflammatory events could have been documented associated with the following conditions: contact lens-induced peripheral ulcer (CLPU), contact lens-induced red eye (CLARE), infiltrative keratitis (IK), asymptomatic infiltrative keratitis (AIK), or asymptomatic infiltrates (AI). These conditions as reported in the studies used in this meta-analysis have been defined by Sweeney et al.38 The second stratum defined a corneal inflammatory event using either an arbitrary author’s definition or a defined severity to qualify as an event. Author’s definitions ranged from CLPU events only,39 peripheral corneal inflammatory events only,40,41 events requiring treatment,42 events resulting in discontinuation from the study24,43 serious and significant corneal inflammatory events,31 or were not defined in the text of the document.22,25,27 Additionally, some studies required the severity of an infiltrative response to be graded on a traditional 4-point ordinal scale. If a defined grade level was required, one study required the defined infiltrative response have a severity equal than or worse than grade 2,20 and one study required the defined infiltrative response have a severity worse than grade 2.28
A generalized linear mixed model (GLMM) framework44 with a random intercept was employed to estimate the overall incidence of corneal inflammatory events from the 23 studies. This modeling framework allows calculation of individual study estimates, adjusting the overall estimate for the effect of covariates, and assessing the heterogeneity across studies. We assumed an underlying Poisson distribution for the occurrence of events. SAS version 9.1 for Windows was used. SAS implements GLMM using the Empirical Bayes method for estimation of the random effects. This approach biases the estimates of studies, in particular those with small sample size, toward the mean estimate of the entire set of studies to deal with their large variance (reduced precision). Hence, an individual estimate utilizing this method is not the observed study estimate, but actually the estimate that would be expected given the overall estimate from all the studies, and the individual study’s events and sample size.
GLMM allows a direct assessment of the heterogeneity across studies based on the significance of the estimate for the variance of the random effects. The latter is the component of variance describing the distribution of the effect across studies. A confidence interval for this random effect variance component including one would indicate the presence of negligible heterogeneity across studies. We could not obtain a null variance for this effect because we were required to define a positive variance value to guarantee the convergence of our models.
Studies were categorized based on whether they reported outcomes with respect to eyes or persons. It was not possible to convert all studies to one uniform unit. In the person-based studies, it was impossible to assess how many eyes were affected. Studies on an eye basis either reported outcomes for both eyes without identifying whether those outcomes occurred in the same patient or without providing a measure of inter-eye correlation to accurately adjust the outcomes to a patient basis. In the case of contralateral eye studies providing eye specific rates, it has been stated that they can be converted to approximate person-based rates if one assumes that the two eyes are independent.45 However, independence is not strictly true when assessing contact lens associated ocular inflammation, and we believe significant error would be introduced using this method. Therefore, we report the results separately depending on whether outcomes were provided on a per-eye or per-person basis. Within each reporting stratum, the significance of the percentage of association between length of wear and material across studies was assessed using the Fischer Exact Test.
Adjusted risk ratios were calculated when possible for covariates where there were enough arms available for various sub-groupings. Although desirable, multivariable adjustment for key covariates was not possible because of the reduced number of studies remaining in each subcategory. Therefore, bivariate adjustments were calculated for each of the following variables: (1) presence of randomization in the study design, (2) location of study, (3) whether older or prototype design SH lenses were used, (4) length of follow-up, (5) definition of infiltrate, and 6) percent of adapted extended lens wearers. In a second subset of analyses, we removed the effect of younger age as a confounding variable46–50 by removing studies where the mean age was <2 standard deviations below the mean age across all studies. Another series of bivariate adjustments were made in studies restricted to the most common lengths of wear by material. Although desirable, we could not adjust for mean refractive error because this covariate was not consistently provided across studies.
We devised a quality scale for evaluating the included studies using nine factors that were most relevant for controlling bias and confounding. Selection and information bias inherent to each study could affect the validity of our results. These biases can occur from differential detection of study outcome and would be best controlled by randomized and prospective designs using both materials. Therefore, a higher quality score was given to randomized or prospective cohort studies studying both SH and low Dk arms. Additionally, information bias can result from inconsistent definitions of study outcome, which can lead to differential disease misclassification. Therefore we stratified and graded all studies based on whether (and what type of) definitions were provided for infiltrative events, as well as whether single events only (vs. recurrent events) could be assumed in the text. Although 16 criteria formed the entire quality assessment, only 9 were considered relevant for further evaluation. Table 1 lists all 16 quality reporting criteria for each study.
We realize that quality scores constructed in such an ad hoc fashion may lack demonstrated validity, therefore, we used the quality scores in evaluations of subgroup analyses rather than using the quality scores as weights in the analysis, per the recommendation of the MOOSE Group.32
Twenty-three studies reporting on 9,336 subjects and 18,537 eyes were selected for analysis. All eligible studies were written in English. Table 2 lists and describes the studies included. There were a total of 31 arms. Eight studies evaluated two arms, but only six (26%) studies directly compared the two materials. Two of the other double-arm studies compared SH lenses at two different lengths of wear. Seventy-eight percent of the studies were prospective and 48% were randomized trials. Five of 15 studies (33%) examining SH lenses used prospective observational or nonrandomized trial designs, whereas 6 of 14 studies (43%) examining low Dk lenses used an observational (five retrospective, one prospective) design. The follow-up periods ranged from 4 to 36 months, the average length of follow-up was 16 months, and the median follow-up was 12 months. Ten studies were performed in the United States, six in Europe, and seven in other countries such as Canada, Australia, India, and Japan. One SH lens study reported using a prototype lens design,29 and seven SH lens studies using lotrafilcon A lenses published before 2003 were assumed to use CIBA Vision’s older lens design and single base curve.
The majority of studies reported outcomes on an eye basis. Fifteen studies (21 arms) encompassing 8,173 subjects or 16,211 eyes reported or allowed accurate calculation on an eye basis, and 8 studies (10 arms) encompassing 1,163 subjects or 2,326 eyes reported outcomes on a person basis. There were six studies providing a direct SH and low Dk lens comparisons, and five of these used an eye based reporting system. For studies reporting on an eye basis, 89% of arms evaluated low Dk lenses on a 7-day regimen, whereas 92% of SH arms evaluated lenses on a 30-day regimen, and this percent of association was highly significant (p = 0.0008). Similarly, for studies reporting on a person basis, 85% of low Dk arms used lenses for 7 days, and 91% of SH arms used lenses for 30 days (p = 0.0006).
Table 3 lists estimates of corneal inflammatory event incidence twice, once for eye-based reporting studies, and again for studies that reported on a person basis. It can be seen in Table 3 that estimates for all grades of infiltrates on a person basis were either not reliable or not provided. Additionally, in the person-based studies, only one study included a direct comparison between SH and low Dk lenses both at 7 days of wear; the remaining seven studies used heterogeneous definitions of infiltrates without direct comparisons between lens types. Therefore valid comparisons between incidence rates and estimates of risk cannot be made in the person based studies and all further results will be based upon eye-based reporting studies only.
To best control for differential disease misclassification bias, the initial analysis of risk utilized a subset of five randomized trials which directly compared the two materials and scored the highest in our quality assessments by containing at least 75% of the predefined key criteria (studies 2, 4, 6, 10, and 22 from Table 2). Of these five studies, four were a contra-lateral randomized eye design, one study randomized patients to bilateral wear of the same material, and all provided event rates on a per-eye basis. Material was perfectly associated with length of wear, that is, all SH arms utilized lenses for up to 30 days, and all low Dk arms utilized lenses for up to 7 days. Table 4 reports the risk ratios for this subset of studies. Overall, the risk ratio summarized from these key studies documents a 2.18 times greater risk of inflammatory events for SH lenses (p < 0.005), and each individual study had approximately a twofold increased risk in our model. The risk ratios correspond to the ratio of the particular event for SH lenses with respect to the low DK lenses. The only study that did not have a significantly elevated risk ratio was study 10 because of high variance in the study’s estimate due to its very small sample size (n = 24).The Forest Plot in Fig. 2 displays the overall risk ratio and study specific risk ratios plus their 95% confidence intervals as determined from our statistical model for these five studies.
Following analysis of the best quality studies, additional arms for both material groups were introduced into subsequent analyses to increase power and to allow for adjustment of potential confounders. The unadjusted risk ratio for SH lenses compared to low Dk lenses for all studies reporting on an eye basis was 2.22 (1.65, 2.98). The heterogeneity in this random intercept model was not significant (p = 0.21), indicating that for this set of studies the overall estimate applies, and no additional sources of variability play a role.
Fig. 3 displays a series of subset analyses where adjusted risk ratios are presented stratified by the adjusting variable and restricted to overall mean age and common length of wear by material. In Figure 3a, unrestricted adjusted risk ratio estimates ranged between 2.18 and 2.23 and all are significant at p ≤ 0.0002. As confounding for age is accounted for by restriction in Figure 3b, risk ratios remained essentially unchanged (all p ≤ 0.0004). Finally, when restricting to common lengths of wear by material (by only including studies where SH lenses were worn for 30 days and low Dk lenses were worn for 7 days) in Figure 3c, the risk ratios again remain unchanged with a range between 2.18 and 2.22 and all are highly significant (p ≤ 0.0007). It was not possible to control for length of wear or disposal schedule because these two covariates were strongly associated with material in such a way that the differences observed between the two materials in the incidence of corneal inflammatory events might be explained in large part by either the length of wear and/or disposal schedule rather than the material itself. It was impossible to disentangle these two potential sources of variability from the effect of material in this analysis. The final analysis in Figure 3c represents perfect confounding between material and length of wear in a manner that portrays the FDA premarket approval studies utilizing these materials.
Contact lens associated corneal inflammatory events are generally considered mild adverse reactions when compared with ulcerative microbial keratitis. Most of these inflammatory events can be treated by simply discontinuing wear, some cases do not require treatment, and some practitioners utilize anti-inflammatory medications to hasten the resolution. The FDA has attempted to use infiltrative keratitis as a surrogate endpoint for microbial keratitis because the very low incidence of the latter would require extremely large and costly premarket studies.45 In all four completed and available 30 day lens premarket approval studies to date,20,21,31,53 significant corneal inflammatory events were a primary safety endpoint. Therefore, the safety of extended wear lens use is judged by many by assessing the occurrence of corneal inflammatory events. To our knowledge, this is the first statistical summary of the literature and unpublished research relating to inflammatory adverse events associated with silicone hydrogel lenses. The results of this meta-analysis are scientifically applicable to researchers as they plan future studies, and clinically applicable to practitioners and subjects that prescribe and wear extended wear contact lenses.
The incidence rates reported here should be considered with respect to the prevalence and/or incidence of corneal infiltrates in noncontact lens wearers. Hickson reported the prevalence of subepithelial microinfiltrates to be 4% of noncontact lens wearing patients57 and Sankaridurg reported 1.6% of eyes had asymptomatic infiltrates at baseline of a longitudinal study enrolling non-contact lens wearing myopes in India.58 Sankaridurg also documented an incidence rate of about 11 asymptomatic events per 100 eye-years of wear in spectacle wearers and about 20 events per 100 eye-years of wear in daily disposable hydrogel wearers.58 These rates are surprisingly high and may be due to their local environment because the cells involved in the infiltrative event have been postulated to mobilize in response to environmental agents such as dust. Alternatively, these high rates may simply be due to meticulous observations and a reporting bias. Nonetheless, it is well documented that extended wear use of contact lenses significantly increases the risk of corneal infiltrates.41,59 In fact, Sankaridurg’s group documented the incidence rate of corneal infiltrative events in a group of subjects wearing low Dk extended wear disposable lenses to be about 44%.36
This meta-analysis reveals that SH lenses typically worn for up to 30 days extended wear probably double the risk for corneal inflammatory events when compared with low Dk lenses when typically worn for 7 days of extended wear. We base this conclusion on the subset of best quality studies which includes five prospective studies that directly compared the two groups of lenses in a randomized fashion and had at least 75% of our best predefined criteria. We also base this increased risk upon subsequent analyses with added arms which increase power and allow the opportunity to control for potential confounders across studies.
This increased risk cannot be definitively linked to SH lens materials because the effect of material on outcome is confounded by length of wear. Multicollinearity would be present in the model if both variables were present in the regression analysis. Therefore, it is not possible to disentangle the increased risk for corneal inflammatory events in SH lens wear from a 30-day wearing schedule. In other words, because we do not have more studies which directly compare and control for length of wear, this meta-analysis cannot address whether length of wear or material is attributable to the development of corneal inflammatory events. Two of our included studies encompassing 658 subjects compared SH lenses at 30 or 7 days of wear.35,43 Both studies reported a higher infiltrative incidence rate in the 30-day wear group; Nilsson35 documented a twofold higher rate (p = 0.09) and Stern et al.,43 using discontinuation from infiltrates as an endpoint, found more than a twofold greater risk. Although Stern et al. found no statistical difference between the groups, a large survival bias was introduced in their study by placing all successfully adapted 30-day SH wearers in the 30-day group. One of our included studies compared both low Dk and SH lenses for 7 days of extended wear.28 This study showed the SH lens to perform clinically and statistically better then the low Dk lens, with a twofold greater risk for infiltrates in the low Dk group. Theoretically, at similar lengths of wear, SH lenses should allow increased flushing of postlens antigenic debris when compared with low Dk lenses.26 It seems plausible that a 30-day nonremoval schedule creates a stagnant tear layer environment which may be more responsible for inflammatory complications than the material itself. The mechanisms may be similar to the development of microbial keratitis, where length of wear has been implicated as the biggest risk factor in low Dk extended wear.50,60
Definitions of adverse events vastly differ among studies. This is not a new problem. In fact, Efron and Morgan have recently shown that differences in case definitions have powerful effects on estimate of risk.61 In our meta-analysis, studies that used the presence/absence of any grade of infiltrate suggest an incidence rate of 7.7 per 100 eye-years for low Dk lenses and 14.4 per 100 eye-years for SH lenses, whereas studies that use some other definition or a required grade to qualify as a significant infiltrate suggest at least a threefold lower incidence rate, regardless of lens material.
A limitation of this study stems from the fact that some studies reporting on an eye basis may have reported on both eyes of a single subject. Because the two eyes are often positively correlated, if the inter-eye correlation is not accounted for in the analysis, the consequence is an overstatement of the precision of the study and a greater measure of statistical significance than the data warrant.62–64 It was not possible to delineate or correct for eye- reported rates in some studies in this meta-analysis. However, because the correlation between the eyes for corneal inflammatory events is not that large, our marginally significant findings may change, but not those which showed a strong statistically significant effect.
Biases across studies can limit the acceptability and appropriateness of data used in a meta-analysis. We have attempted to flag certain biased studies with quality assessments and limit other biases by using broad inclusion criteria. For example, publication bias is always a potential limitation when analyzing the available literature because poor or negative outcomes are not as likely to be reported as are successful or positive results. Additionally, searches of published reports in electronic databases has been shown to find only half of all relevant studies.65 We feel we have minimized this form of bias by including results from five FDA Summaries and three abstracts from major ophthalmic meetings that to our knowledge have not yet been published. Additionally, we have included small case series, single arm cohort studies, or comparative trials where one arm qualified for our analysis because these studies can be used to assess adverse event rates and augment the data from large scale randomized trials.
In summary, SH lenses when worn for 30-day continuous wear have about a twofold greater risk for corneal inflammatory events than low Dk lenses when worn for 7 days. A 30-day wear schedule may predispose the patient to an inflammatory event more than the SH material. Based on the findings of this meta-analysis, how should the clinical community react to prescribing SH lenses for 30 days continuous wear? Based on the vast ocular physiological improvements documented with highly oxygen transmissible SH materials, the relatively low severity of corneal inflammatory events, and the fact that many of these infiltrates are asymptomatic and possibly a normal ocular response to the environment, we continue to promote SH lenses as the lens of choice for extended wear and daily wear. However, a 7-day SH extended wear schedule and a control arm of noncontact lens wearers should be revisited in future studies.
Supported by Research to Prevent Blindness, Ohio Lion’s Eye Research Foundation, and National Eye Institute grants K23 EY015270-01 (to LSF) and R21 EY015145 (to Case Department of Ophthalmology). The work reported in this article was performed while the first author was a recipient of a William C. Ezell Fellowship from the American Optometric Foundation.
LSF has received past research support, unrelated to this study, from CIBAVision, CooperVision, AMO, and Vistakon.
Presented at the 2005 Annual ARVO meeting.
Department of Ophthalmology
Case Western Reserve University
11100 Euclid Ave., Bolwell Bldg. Suite 3200
Cleveland, OH 44106
1. Young C, Horton R. Putting clinical trials into context. Lancet 2005;366:107–8.
2. Papas EB, Vajdic CM, Austen R, Holden BA. High-oxygen-transmissibility soft contact lenses do not induce limbal hyperaemia. Curr Eye Res 1997;16:942–8.
3. Dumbleton K, Richter D, Simpson T, Fonn D. A comparison of the vascular response to extended wear
of conventional lower Dk and experimental high Dk hydrogel contact lenses. Optom Vis Sci 1998;75 (Suppl):170.
4. Keay L, Sweeney DF, Jalbert I, Skotnitsky C, Holden BA. Microcyst response to high Dk/t silicone hydrogel contact lenses
. Optom Vis Sci 2000;77:582–5.
5. Ladage PM, Yamamoto K, Ren DH, Li L, Jester JV, Petroll WM, Cavanagh HD. Effects of rigid and soft contact lens daily wear on corneal epithelium, tear lactate dehydrogenase, and bacterial binding to exfoliated epithelial cells. Ophthalmology 2001;108:1279–88.
6. Lin MC, Graham AD, Polse KA, McNamara NA, Tieu TG. The effects of one-hour wear of high-Dk soft contact lenses on corneal pH and epithelial permeability. CLAO J 2000;26:130–3.
7. Doughty MJ, Aakre BM, Ystenaes AE, Svarverud E. Short-term adaptation of the human corneal endothelium to continuous wear of silicone hydrogel (lotrafilcon A) contact lenses after daily hydrogel lens wear. Optom Vis Sci 2005;82:473–80.
8. Latkovic S, Nilsson SE. The effect of high and low Dk/L soft contact lenses on the glycocalyx layer of the corneal epithelium and on the membrane associated receptors for lectins. CLAO J 1997;23:185–91.
9. Ren DH, Petroll WM, Jester JV, Ho-Fan J, Cavanagh HD. The relationship between contact lens oxygen permeability and binding of Pseudomonas aeruginosa
to human corneal epithelial cells after overnight and extended wear
. CLAO J 1999;25:80–100.
10. Ren DH, Petroll WM, Jester JV, Ho-Fan J, Cavanagh HD. Short-term hypoxia downregulates epithelial cell desquamation in vivo, but does not increase Pseudomonas aeruginosa
adherence to exfoliated human corneal epithelial cells. CLAO J 1999;25:73–9.
11. Ren DH, Yamamoto K, Ladage PM, Molai M, Li L, Petroll WM, Jester JV, Cavanagh HD. Adaptive effects of 30-night wear of hyper-O(2) transmissible contact lenses on bacterial binding and corneal epithelium: a 1-year clinical trial. Ophthalmology 2002;109:27–39.
12. Imayasu M, Petroll WM, Jester JV, Patel SK, Ohashi J, Cavanagh HD. The relation between contact lens oxygen transmissibility and binding of Pseudomonas aeruginosa
to the cornea after overnight wear. Ophthalmology 1994;101:371–88.
13. Ren H, Petroll WM, Jester JV, Cavanagh HD, Mathers WD, Bonanno JA, Kennedy RH. Adherence of Pseudomonas aeruginosa
to shed rabbit corneal epithelial cells after overnight wear of contact lenses. CLAO J 1997;23:63–8.
14. Schein OD, McNally JJ, Katz J, Chalmers RL, Tielsch JM, Alfonso E, Bullimore M, O’Day D, Shovlin J. The incidence of microbial keratitis among wearers of a 30-day silicone hydrogel extended-wear contact lens. Ophthalmology 2005;112:2172–9.
15. Cheng KH, Leung SL, Hoekman HW, Beekhuis WH, Mulder PG, Geerards AJ, Kijlstra A. Incidence of contact-lens-associated microbial keratitis and its related morbidity. Lancet 1999;354:181–5.
16. Poggio EC, Glynn RJ, Schein OD, Seddon JM, Shannon MJ, Scardino VA, Kenyon KR. The incidence of ulcerative keratitis among users of daily-wear and extended-wear soft contact lenses. N Engl J Med 1989;321:779–83.
17. Stapleton F, Keay L, Edwards K, Naduvilath T, Dart J, Brian G, Holden BA. Risks of contact lens related microbial keratitis. Optom Vis Sci 2005;82:E-abstract 050068.
18. Morgan PB, Efron N, Hill EA, Raynor MK, Whiting MA, Tullo AB. Incidence of keratitis of varying severity among contact lens wearers. Br J Ophthalmol 2005;89:430–6.
19. Efron N, Morgan PB, Hill EA, Raynor MK, Tullo AB. Incidence and morbidity of hospital-presenting corneal infiltrative events associated with contact lens wear. Clin Exp Optom 2005;88:232–9.
20. U.S. Food and Drug Administration. Premarket Application Summary of Safety and Effectiveness Data. PureVision visibility tinted contact lens. November 20, 2001. Available at: www.fda.gov/cdrh/pdf/P980006S004b.doc
. Accessed Janaury 19, 2007.
21. U.S. Food and Drug Administration. Summary of Safety and Effectiveness. CIBAVision Focus Night and Day (lotrafilcon A) soft contact lens. October 12, 2001. Available at: www.fda.gov/cdrh/pdf/p000030b.pdf
. Accessed January 19, 2007.
22. Dumbleton K, Fonn D, Jones L, Williams-Lyn D, Richter D. Severity and management of contact lens related complications with continuous wear of high Dk silicone hydrogel lenses. Optom Vis Sci 2000;77 (Suppl):216.
23. Iruzubieta JM, Ripoll JRN, Chiva J, Fernandez OE, Alvarez JJR, Delgado F, Villa C, Traverso LM. Practical experience with a high Dk lotrafilcon A fluorosilicone hydrogel extended wear
contact lens in Spain. CLAO J 2001;27:41–6.
24. Morgan PB, Efron N, Maldonado-Codina C, Efron S. Adverse events and discontinuations with rigid and soft hyper Dk contact lenses used for continuous wear. Optom Vis Sci 2005;82:528–35.
25. Sweeney DF, Stern J, Naduvilath T, Holden BA. Inflammatory adverse event rates over 3 years with silicone hydrogel lenses. Invest Ophthalmol Vis Sci 2002;43:E-abstract 976.
26. Paugh JR, Stapleton F, Keay L, Ho A. Tear exchange under hydrogel contact lenses: methodological considerations. Invest Ophthalmol Vis Sci 2001;42:2813–20.
27. Brennan NA, Coles ML, Comstock TL, Levy B. A 1-year prospective clinical trial of balafilcon a (PureVision) silicone-hydrogel contact lenses used on a 30-day continuous wear schedule. Ophthalmology 2002;109:1172–7.
28. U.S. Food and Drug Administration. 7-Day Extended Wear
of CIBAVision Focus Night and Day (lotrafilcon A) soft contact lens. October 12, 2001. Available at:www.fda.gov/cdrh/pdf/p000030b.pdf
. Accessed January 19, 2007.
29. Fonn D, MacDonald KE, Richter D, Pritchard N. The ocular response to extended wear
of a high Dk silicone hydrogel contact lens. Clin Exp Optom 2002;85:176–82.
30. Malet F, Pagot R, Peyre C, Subirana X, Lejeune S, George-Vicariot MN, Bleshoy H, Long B. Clinical results comparing high-oxygen and low-oxygen permeable soft contact lenses in France. Eye Contact Lens 2003;29:50–4.
31. U.S. Food and Drug Administration. Premarket Application Summary of Safety and Effectiveness Data. Vistakon (senofilcon A) Contact Lens. August 20, 2004. Available at: www.fda.gov/cdrh/pdf4/k042275.pdf
. Accessed January 19, 2007.
32. Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, Moher D, Becker BJ, Sipe TA, Thacker SB. Meta-analysis
of observational studies in epidemiology: a proposal for reporting. Meta-analysis
of Observational Studies in Epidemiology (MOOSE) group. JAMA 2000;283:2008–12.
33. Gleason W, Tanaka H, Albright RA, Cavanagh HD. A 1-year prospective clinical trial of menicon Z (tisilfocon A) rigid gas-permeable contact lenses worn on a 30-day continuous wear schedule. Eye Contact Lens 2003;29:2–9.
34. Morgan PB, Efron N. Comparative clinical performance of two silicone hydrogel contact lenses
for continuous wear. Clin Exp Optom 2002;85:183–92.
35. Nilsson SE. Seven-day extended wear
and 30-day continuous wear of high oxygen transmissibility soft silicone hydrogel contact lenses
: a randomized 1-year study of 504 patients. CLAO J 2001;27:125–36.
36. Sankaridurg PR, Sweeney DF, Sharma S, Gora R, Naduvilath T, Ramachandran L, Holden BA, Rao GN. Adverse events with extended wear
of disposable hydrogels: results for the first 13 months of lens wear. Ophthalmology 1999;106:1671–80.
37. Szczotka-Flynn L, Debanne S, Cheruvu V, Long B, Dillehay S, Barr J, Bergenske P, Donshik P, Secor G, Yoakum J. Predictive factors for corneal infiltrates
with continuous wear of silicone hydrogel contact lenses
. Arch Ophthalmol, in press.
38. Sweeney DF, Jalbert I, Covey M, Sankaridurg PR, Vajdic C, Holden BA, Sharma S, Ramachandran L, Willcox MD, Rao GN. Clinical characterization of corneal infiltrative events observed with soft contact lens wear. Cornea 2003;22:435–42.
39. Aakre BM, Ystenaes AE, Doughty MJ, Austrheim O, Westerfjell B, Lie MT. A 6-month follow-up of successful refits from daily disposable soft contact lenses to continuous wear of high-Dk silicone-hydrogel lenses. Ophthalmic Physiol Opt 2004;24:130–41.
40. Boswall GJ, Ehlers WH, Luistro A, Worrall M, Donshik PC. A comparison of conventional and disposable extended wear
contact lenses. CLAO J 1993;19:158–65.
41. Levy B, McNamara N, Corzine J, Abbott RL. Prospective trial of daily and extended wear
disposable contact lenses. Cornea 1997;16:274–6.
42. Maguen E, Tsai JC, Martinez M, Rosner I, Cot PC, Macy JI, Nesburn AB. A retrospective study of disposable extended-wear lenses in 100 patients. Ophthalmology 1991;98:1685–9.
43. Stern J, Wong R, Naduvilath TJ, Stretton S, Holden BA, Sweeney DF. Comparison of the performance of 6- or 30-night extended wear
schedules with silicone hydrogel lenses over 3 years. Optom Vis Sci 2004;81:398–406.
44. Platt RW, Leroux BG, Breslow N. Generalized linear mixed models for meta-analysis
. Stat Med 1999;18:643–54.
45. Saviola JF, Hilmantel G, Rosenthal AR. The U. S. Food and Drug Administration’s role in contact lens development and safety. Eye Contact Lens 2003;29:S160−S165.
46. Brennan NA, Mullen B. Increased susceptibility of younger people to adverse reactions with silicone-hydrogel contact lens continuous wear. Optom Vis Sci 2001;78 (Suppl):229.
47. Chalmers RL, McNally J, Schein O, Katz J, Tielsch J. Rick factors for infiltrates in continuous wear contact lenses (Abstract). Optom Vis Sci 2005;82:E-abstract 050076.
48. du Toit R, John T, Sweeney DF, Kalliris A, Tahhan N, Papas E. Association of age, gender and ethnicity with the incidence of adverse responses with extended wear
of silicone hydrogel lenses. Optom Vis Sci 2002;79 (Suppl):9.
49. McNally JJ, Chalmers RL, McKenney CD, Robirds S. Risk factors for corneal infiltrative events with 30-night continuous wear of silicone hydrogel lenses. Eye Contact Lens 2003;29:S153−S156.
50. Stapleton F, Dart JK, Minassian D. Risk factors with contact lens related suppurative keratitis. CLAO J 1993;19:204–10.
51. Buckley CA, Buckley CJ, Griffiths J. Extended wear
disposable soft contact lenses as an alternative to photorefractive keratectomy: report of 4 years experience. Aust NZ J Ophthalmol 1997;25:111–16.
52. Dillehay S, Long B, Barr J, Bergenske P, Donshik P, Secor G, Yoakum J. Patient experience and management in hDk silicone hydrogel soft contact lenses. Optom Vis Sci 2003;80 (Suppl):186.
53. U.S. Food and Drug Administration. Summary of Safety and Effectiveness. Menicon Z (tisilfocon A) rigid gas permeable (RGP) contact lens. July 12, 2002. Available at: www.fda.gov/cdrh/pdf/P990018S002b.pdf
. Accessed January 2007;19.
54. Maguen E, Rosner I, Caroline P, Macy JI, Nesburn AB. A retrospective study of disposable extended wear
lenses in 100 patients: year 2. CLAO J 1992;18:229–31.
55. Poggio EC, Abelson M. Complications and symptoms in disposable extended wear
lenses compared with conventional soft daily wear and soft extended wear
lenses. CLAO J 1993;19:31–9.
56. Watanabe K, Hamano H. The typical pattern of superficial punctate keratopathy in wearers of extended wear
disposable contact lenses. CLAO J 1997;23:134–7.
57. Hickson S, Papas E. Prevalence of idiopathic corneal anomalies in a non contact lens-wearing population. Optom Vis Sci 1997;74:293–7.
58. Sankaridurg PR, Sweeney DF, Holden BA, Naduvilath T, Velala I, Gora R, Krishnamachary M, Rao GN. Comparison of adverse events with daily disposable hydrogels and spectacle wear: results from a 12-month prospective clinical trial. Ophthalmology 2003;110:2327–34.
59. Cutter GR, Chalmers RL, Roseman M. The clinical presentation, prevalence, and risk factors of focal corneal infiltrates
in soft contact lens wearers. CLAO J 1996;22:30–7.
60. Dart JK, Stapleton F, Minassian D. Contact lenses and other risk factors in microbial keratitis. Lancet 1991;338:650–3.
61. Efron N, Morgan PB. Impact of differences in diagnostic criteria when determining the incidence of contact lens-associated keratitis. Optom Vis Sci 2006;83:152–9.
62. Katz J. Two eyes or one? The data analyst’s dilemma. Ophthalmic Surg 1988;19:585–9.
63. Katz J, Zeger S, Liang KY. Appropriate statistical methods to account for similarities in binary outcomes between fellow eyes [see comments]. Invest Ophthalmol Vis Sci 1994;35:2461–5.
64. Ray WA, O’Day DM. Statistical analysis of multi-eye data in ophthalmic research. Invest Ophthalmol Vis Sci 1985;26:1186–8.
65. McManus RJ, Wilson S, Delaney BC, Fitzmaurice DA, Hyde CJ, Tobias RS, Jowett S, Hobbs FD. Review of the usefulness of contacting other experts when conducting a literature search for systematic reviews. BMJ 1998;317:1562–3.