(Investigative Ophthalmology and Visual Science. 2005;46:3136-3143.)
© 2005 by The Association for Research in Vision and Ophthalmology, Inc.
DOI: 10.1167/iovs.05-0133
Risk Factors for the Development of Corneal Infiltrative Events Associated with Contact Lens Wear
Philip B. Morgan,1
Nathan Efron,1
Noel A. Brennan,2
Elizabeth A. Hill,1
Mathew K. Raynor,3 and
Andrew B. Tullo3
1From Eurolens Research, The University of Manchester, Manchester, United Kingdom;
2Brennan Consultants, Melbourne, Australia; and the
3Royal Eye Hospital, Manchester, United Kingdom.
 |
Abstract
|
|---|
PURPOSE. To identify risk factors for the development of corneal infiltrative events (CIEs) associated with contact lens wear, and to report other relevant clinical characteristics.
METHODS. A series of symptomatic contact lens wearers presenting consecutively to a large hospital clinic over a 1-year period were examined. The clinical severity of any CIE was determined with a scoring system, and general and lens-specific information was captured with a questionnaire. Logistic regression analyses were performed to investigate the association between a range of risk factors and the occurrence of any CIE and the subset of cases categorized as severe keratitis. Three quasi-independent control groups were used for this analysis: hospital, lens-matched, and population. The relationship between clinical severity and the delay in attending the hospital was investigated. The prevalence of symptoms and initial actions taken by patients are reported.
RESULTS. Factors identified as being associated with an increased risk of development of a CIE include: wearing modality/lens type (greatest risk for extended-wear hydrogel lenses of 7.1 vs. daily wear hydrogel lenses), male gender (relative risk 1.4), smoking (1.4), the absence of relevant ocular (1.8) and general health (2.4) problems, and the late winter months (greatest risk in March of 3.6 vs. July). The overall predictive value of these risk factors for a given individual was low. Shorter time delays in hospital attendance were associated with increasing severity of keratitis. Eye soreness was the most common initial symptom (prevalence 69%), and the most frequent initial course of action taken by the patient was lens removal (prevalence 50%).
CONCLUSIONS. Risk factors for the development of contact lens keratitis were identified that, although being of limited predictive value for individual patients, highlight general associations that may assist in the management of contact lens wearers. Such risk factors may also assist in the development of a more complete understanding of the etiology of contact-lens–associated CEIs.
In a recent paper,1 we described the incidence of severe and nonsevere keratitis among patients wearing all current forms of contact lenses designed for both daily and extended wear. We adopted the novel approach of considering contact lens–associated corneal infiltrative events (CIEs) as a continuous spectrum of disease severity. This approach represents an alterative to the more traditional one of attempting to differentiate between so-called microbial and sterile keratitis.2 3 4 A CIE is defined as a clinical event characterized by an aggregation of inflammatory cells in the cornea, irrespective of etiology.3
The identification of the type and magnitude of risk in relation to the development of CIEs is a useful adjunct to clinical practice if the risk factors are capable of predicting outcomes. Numerous investigators have identified risk factors for the development of either sterile or microbial keratitis for daily and extended contact lens wear (Tables 1 2 3) . Few authors have addressed the question of risk of CIEs in association with silicone hydrogel lens wear. McNally et al.6 reported that those who wear silicone hydrogel lenses on an extended-wear basis have an increased risk of infiltrative events other than microbial keratitis if they are under the age of 30 years. The risk increases if they smoke tobacco.
As part of the design of our earlier study,1 we collected information on a range of factors ancillary to lens type (including silicone hydrogel lenses) and modality of wear that may have had an influence on the incidence of the full spectrum of CIEs. The purpose of this study was to identify which of these factors show significant associations with the development of CIEs and to consider possible clinical implications.
|
Methods
|
|---|
A survey was undertaken of contact lens wearers consecutively attending the acute service of the Royal Eye Hospital (Manchester, UK), irrespective of their reason for coming to the hospital, between January 25, 2003, and January 24, 2004. The study patients were asked to complete a patient survey questionnaire to provide information pertaining to their contact-lens–wearing experience and circumstances relating to their coming to the hospital.
Table 4 lists the variables extracted from the questionnaire that were entered into the analysis. Occupation was scored on a scale from 1 to 10 by three of us, according to our perception of socioeconomic standing based on the description provided by the patient. A score of 10 corresponded to an occupation of a highly professional nature with strong intellectual demands. A score of 1 corresponded to an occupation of an unskilled and/or menial nature with minimal intellectual demands. Health problems deemed to be relevant are those that were judged by three of us to have implications with respect to contact lens wear. For example, chronic abdominal pain was considered to be unrelated to lens wear, whereas diabetes was deemed to be of relevance. Eye problems thought to be relevant are those that were judged by three of us to have implications with respect to contact lens wear. For example, retinal detachment was judged to be unrelated to lens wear, whereas dry eye was taken to be of relevance. Seasonal influence was investigated by considering the month in which the patients attended the hospital.
Because of the small number of patients in certain categories of lens wear, we classified wearing modality/lens type for the purposes of this analysis into the following five groups: daily wear rigid lenses; daily wear hydrogel daily disposable lenses; daily wear hydrogel lenses (i.e., all daily wear hydrogel lenses excluding daily disposable lenses); extended wear hydrogel lenses (including daily disposable lenses that were used on an extended-wear basis in a noncompliant manner); and extended-wear silicone hydrogel lenses.
For each study patient examined, the attending clinician made a judgment as to whether there was any form of CIE, which may or may not have been associated with corneal ulceration. If there was no evidence of a CIE, no further information was captured for the purposes of this study. If there was a CIE, the attending clinician completed a clinical survey form (on the reverse side of the patient survey questionnaire) which enabled the severity of the CIE to be assigned a score (from 2 to 22) based on a clinical severity matrix described by Aasuri et al.17 and subsequently modified by us.1 For the purposes of discussion, these patients are referred to as the CIE group. For certain aspects of our analysis, cases of severe keratitis were identified as those having a clinical severity score >8.
Three quasi-independent control groups were used in the analysis, the justification for which is considered in the Discussion section. First, data relating to patients presenting to the hospital who were contact lens wearers but did not have any form of CIE formed the basis of a hospital control group. Second, we sought to create a matched control group of contact lens wearers without CIEs, consisting of the same number of patients in each of the wearing modality/lens type groups as that found in the CIE group. This control group was assembled by recruiting patients from the Eurolens Research subject database whose records of contact lens wear matched the required wearing modality/lens type criteria. Third, the control described in our previous publication1 constituted a population control group, which comprised projections of the number of those in the hospital catchment population who were prescribed the various wearing modality/lens types. Because of the distinct nature of each of these control groups, different variables were available for entry into the logistic regression models (see Table 4 and the description of the statistical analysis in the following text for more details).
Descriptive statistics relating to features of interest within the CIE group were collated. Further analyses that did not require a control group, such as correlation between factors relating to patients who had CIEs, were undertaken in an attempt to define associations of clinical interest.
The research adhered to the tenets of the Declaration of Helsinki. Informed consent was obtained from all patients surveyed in this study. Ethics approval to conduct the study was obtained from the Local Research Ethics Committee (Central) of Manchester Health Authority.
A series of stepwise logistic regression models (JMP; SAS Institute, Inc., Cary, NC) were constructed using possible risk factors as the independent variables and the binary outcome of the presence or absence of CIEs as the dependent variable. This series of models comprised an analysis for two categories of CIE (all CIEs and severe keratitis) tested against the three control groups (hospital, matched, and population), giving a total of six separate logistic regression models.
The logistic regression models enabled estimates of statistically significant odds ratios; however, for ease of data interpretation, these ratios were converted to relative risks with appropriate 95% confidence intervals. Statistical outcomes generating P < 0.05 were considered to be statistically significant and probabilities 0.05 and 0.10 were considered to be weakly significant. The statistic "lack of fit" was assessed as a precaution against overmodeling of the data. The strength of the overall associations accounted for by the model was assessed by considering pseudo-r2 values generated by the software.
|
Results
|
|---|
A total of 415 contact lens wearers attended the study center over the 12-month survey period. Useable responses to the patient survey questionnaire were obtained from 410 of these lens wearers. Of these, 118 were observed to have a CIE, and a clinical survey form was completed by the attending clinician for each of these patients. The remaining 292 patients who completed the survey questionnaire forms constituted the hospital control group. The matched control group comprised 116 subjects. The number of population controls has been presented previously.1 Table 5 gives a breakdown of the number of subjects in the CIE group and the three control groups, stratified for wearing modality/lens type.
Table 6 presents a matrix of the relative risk, 95% confidence intervals and probabilities for levels of the significant factors as identified in the six logistic regression models. Essentially, each of the columns with numerical data represents the result of a separate logistic regression. As can be seen from this table, the following six factors were identified as statistically significant in at least one of the logistic regression models: wearing modality/lens type, gender, smoking, relevant eye problem, relevant health problem, and month of year. A comparison of particular interest is the risk of developing CIEs while wearing hydrogel lenses versus silicone hydrogel lenses on an extended-wear basis. In the hospital control group, there was a higher risk of severe keratitis in the extended-wear hydrogel group versus the extended-wear silicone hydrogel group (relative risk; 95% CI: 3.78; 1.07–13.35) but no difference for all CIEs (1.28; 0.71–2.33). For the population control, there was also a higher risk of severe keratitis in the extended wear hydrogel group versus the extended-wear silicone hydrogel group (4.87; 1.10–21.69) but again, there was no difference for all CIEs (0.82; 0.33–2.05).
View this table:
[in this window]
[in a new window]
|
TABLE 6. Relative Risk, 95% CI, and Probabilities for Various Levels of Significant Factors for the Logistic Regression Models
|
|
The number of contact lens wearers attending the hospital clinic with nonsevere keratitis, severe keratitis, and other ocular problems is displayed in Figure 1 for the 12 months of the study survey. Also plotted on this graph, for interest, is the number of consultations for influenza-like illness to the UK National Health Service Helpline per 100,000 persons aged 15 to 64 in England during this same period. The significance of this latter plot is considered in the Discussion.
View larger version (26K):
[in this window]
[in a new window]
|
FIGURE 1. Proportion of contact-lens–wearing patients attending the acute service of the Royal Eye Hospital, Manchester, during the study period, with severe keratitis, nonsevere keratitis, and other problems. Also shown is the number of consultations for influenza-like illness to the telephone Helpline of the UK National Health Service (NHS) per 100,000 persons in England aged 15 to 64, over the same period.
|
|
Tables 7 and 8 describe the percentage frequency of reported symptoms and the immediate course of action taken, respectively, among the CIE group. A higher severity of CIE was significantly associated with a shorter delay between the onset of symptoms and attendance at the hospital (Spearman Rank Correlation, z [corrected for ties] = –2.7, P < 0.01); this relationship is illustrated in Figure 2 .
View larger version (11K):
[in this window]
[in a new window]
|
FIGURE 2. Delay in attending the acute service of the Royal Eye Hospital, Manchester, from the initial onset of symptoms. Larger circles: multiple patients.
|
|
|
Discussion
|
|---|
Obtaining a representative control group in epidemiologic studies usually presents a considerable challenge. In this work, three quasi-independent control groups were used. Use of the hospital control group was based on the assumption that attendance at the hospital was independent of the risk factors associated with contact lens wear. Of course, this control group of contact lens wearers reported to the hospital because they were symptomatic, and as such, they may not represent an accurate cross section of contact lens wearers within the population as a whole. We believe that a disproportionately high number of extended-wear lens users may be represented in this control group because of the greater physiological impact of this modality of lens wear; however, the proportions of the different lens types (i.e., rigid versus daily disposables versus hydrogels versus silicone hydrogels) probably represent a reasonable reflection of lens wear within the general community.
We also assembled an external control group of contact lens wearers, matched for wearing modality/lens type, who had not had a CIE. A limitation of this matched control group is that it restricted assessment to risk factors other than wearing modality/lens type and seasonal effects. In addition, the formation of a matched control group from a different demographic population (in this case, subjects participating in university-based contact lens clinical trials) may have influenced the results.
The population control, which was used to derive incidence figures for CIEs in our previous study,1 essentially constituted a third control group. These data consisted of projections of the number of wearers in the hospital catchment population based on earlier survey data relating to prescribing habits of wearing modality/lens type. Therefore, information was not available in relation to factors such as age, gender, smoking, and relevant eye and health problems, which limited the number of risk factors that could be analyzed in the relevant logistic regressions. A further limitation of the population control group is that actual patterns of lens wear in the population may not be in precise concordance with what was prescribed. Because the estimate of prescribed extended wear in hydrogel lens wearers is low, the relative risk of a CIE associated with this form of lens wear is sensitive to small fluctuations in this estimate. We were therefore interested in observing that the statistically significant finding of higher risk for severe keratitis associated with extended-wear hydrogel lenses compared with that of extended-wear silicone hydrogel lenses found in the population control (relative risk, x4.87) was replicated when using the hospital control (x3.78), notwithstanding the potential lack of robustness of both of these control groups.
Several observations can be made regarding the risks of CIEs with various wearing modalities and lens types. Irrespective of the control group, daily wear of rigid lenses was found to cause a lower risk of CIEs compared with daily wear of hydrogel lenses. Sleeping in hydrogel and silicone hydrogel contact lenses was associated with a higher risk of CIEs than was daily hydrogel lens wear. These findings are consistent with the general sentiment expressed by numerous investigators of the risks of sleeping in contact lenses (see Table 6 ).
Males have a 30% to 40% increased risk of the occurrence of CIEs compared with females. This gender difference has been reported previously for microbial keratitis.4 8 The reason for this association is not obvious, but it may be related to perceived health risks. Males have different attitudes and perceptions relating to health risks than do females, in that they perceive risks as much smaller and much more acceptable.18 For example, males may be more inclined to undervalue the importance of personal hygiene with respect to contact lens care.
Smoking was found to be associated with a 35% greater risk of CIEs, and this was increased to twofold for severe keratitis. Others have reported similar findings.2 7 Smoking may be a risk factor for several reasons. It is generally considered that smoking is an unhygienic pursuit, which may be linked to a general lack of hygiene with respect to matters relating to contact lens wear and care. Toxins from smoke may either irritate the eyes directly19 or become absorbed into the contact lens an act as an irritant that compromises the health of the ocular surface20 and predisposes the eye to a CIE. Smoking is known to have an immunomodulatory effect, and in that way cigarette smokers may be predisposed to a CIE.21
There is approximately a two times greater risk of a CIE in the absence of compromised ocular and general health. The protective effect of compromised ocular health in lowering the risk of contact lens–associated CIEs may be explained by the precautionary attitude adopted by those with compromised ocular health, in that such persons may, for example, cease lens wear, reduce wearing time, or use self-prescribed topical ocular medications to alleviate the condition. Such strategies may have the secondary effect of precluding a CIE. An alternative explanation is that compromised ocular health may be associated with a general upregulation of the innate defense status of the eye, so that there is an ever-present resistance to extraneous challenges to the ocular surface which could result in a CIE. These principles can be extended to explain why compromised general health also serves to protect the eye from a CIE.
The notion that adverse ocular conditions related to contact lens wear can be influenced by the time of year (seasonal effect) is well established; for example, Begley et al.22 reported that the onset of contact lens–associated papillary conjunctivitis was seasonal, in that the incidence of this condition peaked during the allergy seasons in the midwestern United States. It may be supposed in the present study that the incidence of CIEs is related to seasonal variations in ocular or general health. We found a two to fourfold increased risk of CIEs in late winter (January to March) in the United Kingdom, compared with the risk in midsummer (July). We accessed the number of consultations for influenza-like illness to the UK National Health Service Helpline by people aged 15 to 64 years in England during the same time period as our study and found that this number peaked around October and November 2003,23 which is in discordance with the peak incidence of CIEs in our study from January to March 2003 (Fig. 1) . Of interest, this observation is consistent with the finding described earlier of a lower incidence of CIEs in association with compromised general health.
A noteworthy statistic in our logistic regression analyses is the pseudo-r2 values. These are analogous to the r2 values generated in conventional linear regression models, which indicate the amount of variance within the dataset that can be accounted for by the identified significant factors. For the six regression analyses reported herein, the pseudo-r2 values were all less than 10%. This suggests that any given risk factor (or, indeed, group of risk factors) is of limited value in predicting the likelihood of a given individual’s having a CIE, notwithstanding the general associations that we have demonstrated. Although numerous investigators have also identified risk factors associated with keratitis among contact lens wearers (Tables 1 2 3) , they have failed to report the strength of those associations. If such associations are weak, then the predictive capacity of the risk factors for an individual patient is potentially exaggerated. In view of the benefit of knowing these associations, we suggest that future studies would be enhanced by citing r2 or pseudo-r2 values.
In our analysis, several of the factors listed in Table 4 were not found to be statistically significant—namely: age, occupation, contact lens care solutions, and swimming. Our failure to establish significance is due to either (1) there being insufficient power in our dataset to enable a demonstration of statistical significance or (2) there being sufficient power, but the factor was genuinely not associated with the development of a CIE. It is for these reasons that we have been unable to verify the findings of previous studies of an association of age,6 occupation,4 contact lens care solutions,16 and swimming15 with CIEs among contact lens wearers.
It is not surprising that ocular discomfort was the symptom that was first noticed by patients who had CIEs. This symptom was reported by 69% of patients. The other prominent symptom noticed at onset was eye redness. These onset symptoms are similar to those previously reported by patients who have nonsevere3 or severe2 keratitis.
From a clinical perspective, it is comforting to note that the most common immediate course of action for a patient who experiences a CIE is to remove the contact lenses, and that the second most common action is to seek professional advice. Patients consulted a variety of professionals, including optometrists, opticians, general medical practitioners, pharmacists, and emergency hospital staff. Some of the more unusual courses of immediate action included going to sleep and placing cold cucumber slices or moist tea bags on the closed eye.
We found an inverse relationship between the clinical severity score and the time of delay in attending the hospital. It is interesting that there is considerable variance in the data. For example, three patients with clinical severity scores >8 (severe keratitis) delayed their attendance at the hospital by more than 1 week. It may be that these patients had only mild symptoms in the first week and their symptoms became more severe immediately before attending the hospital, which would explain their delayed attendance. In general, if it is assumed that the level of ocular discomfort is similar irrespective of the delay in attending the hospital (due to a pain-intolerance threshold having been crossed), then a more severe condition that might be expected to progress more rapidly is associated with a shorter delay in attending the hospital. Conversely, a more slowly progressing condition of lower severity might be associated with a longer delay in attending the hospital.
When dealing with CIEs in practice, contact lens clinicians may wish to consider the information presented in this article relating to risk factors, reported symptoms, actions taken by patients, and their delay in attending the hospital. Factors identified as being associated with an increased risk of a CIE include wearing modality/lens type (greatest risk for extended wear hydrogel lenses), being male, smoking, the absence of relevant ocular and general health problems, and the late winter months. As highlighted in this article, such risk factors may also assist in the development of a more complete understanding of the etiology of contact lens–associated corneal infiltrative events.
|
Acknowledgements
|
|---|
The authors thank colleagues who work in the acute service of the Royal Eye Hospital—in particular, John Uddin, and Linda Raynor—for their enthusiastic and conscientious participation in this survey.
|
Footnotes
|
|---|
Submitted for publication February 1, 2005; revised May 26, 2005; accepted July 20, 2005.
Disclosure: P.B. Morgan, None; N. Efron, None; N.A. Brennan, None; E.A. Hill, None; M.K. Raynor, None; A.B. Tullo, None
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked "advertisement" in accordance with 18 U.S.C. 
1734 solely to indicate this fact.
Corresponding author: Philip B. Morgan, Eurolens Research, The University of Manchester, Manchester M60 1QD, UK; philip.morgan{at}manchester.ac.uk.
|
References
|
|---|
- 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;89:430–436.[Abstract/Free Full Text]
- Lam DS, Houang E, Fan DS, Lyon D, Seal D, Wong E. Incidence and risk factors for microbial keratitis in Hong Kong: comparison with Europe and North America. Eye. 2002;16:608–618.[CrossRef][ISI][Medline][Order article via Infotrieve]
- Sweeney DF, Jalbert I, Covey M, et al. Clinical characterization of corneal infiltrative events observed with soft contact lens wear. Cornea. 2003;22:435–442.[CrossRef][ISI][Medline][Order article via Infotrieve]
- Dart JK, Stapleton F, Minassian D. Contact lenses and other risk factors in microbial keratitis. Lancet. 1991;338:650–653.[CrossRef][ISI][Medline][Order article via Infotrieve]
- Cutter GR, Chalmers RL, Roseman M. The clinical presentation, prevalence, and risk factors of focal corneal infiltrates in soft contact lens wearers. Contact Lens Assoc Ophthalmol J. 1996;22:30–37.
- 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–S166.[Medline][Order article via Infotrieve]
- Schein OD, Glynn RJ, Poggio EC, Seddon JM, Kenyon KR. The relative risk of ulcerative keratitis among users of daily-wear and extended-wear soft contact lenses: a case-control study. Microbial Keratitis Study Group. N Engl J Med. 1989;321:773–778.[Abstract]
- Poggio EC, Glynn RJ, Schein OD, et al. The incidence of ulcerative keratitis among users of daily-wear and extended-wear soft contact lenses. N Engl J Med. 1989;321:779–783.[Abstract]
- Stapleton F, Dart JK, Minassian D. Risk factors with contact lens related suppurative keratitis. Contact Lens Assoc Ophthalmol J. 1993;19:204–210.
- MacRae S, Herman C, Stulting RD, et al. Corneal ulcer and adverse reaction in premarket contact lens studies. Am J Ophthalmol. 1991;111:457–465.[ISI][Medline][Order article via Infotrieve]
- Matthews TD, Frazer DG, Minassian DC, Radford CF, Dart JK. Risks of keratitis and patterns of use with disposable contact lenses. Arch Ophthalmol. 1992;110:1559–1562.[Abstract]
- Buehler PO, Schein OD, Stamler JF, Verdier DD, Katz J. The increased risk of ulcerative keratitis among disposable soft contact lens users. Arch Ophthalmol. 1992;110:1555–1558.[Abstract]
- Nilsson SE, Montan PG. The hospitalized cases of contact lens induced keratitis in Sweden and their relation to lens type and wear schedule: results of a three-year retrospective study. Contact Lens Assoc Ophthalmol J. 1994;20:97–101.
- Nilsson SE, Montan PG. The annualized incidence of contact lens induced keratitis in Sweden and its relation to lens type and wear schedule: results of a 3-month prospective study. Contact Lens Assoc Ophthalmol J. 1994;20:225–230.
- Radford CF, Minassian DC, Dart JK. Acanthamoeba keratitis in England and Wales: incidence, outcome, and risk factors. Br J Ophthalmol. 2002;86:536–542.[Abstract/Free Full Text]
- Radford CF, Bacon AS, Dart JK, Minassian DC. Risk factors for acanthamoeba keratitis in contact lens users: a case-control study. BMJ. 1995;310:1567–1570.[Abstract/Free Full Text]
- Aasuri MK, Venkata N, Kumar VM. Differential diagnosis of microbial keratitis and contact lens-induced peripheral ulcer. Eye Contact Lens. 2003;29:S60–S62.[Medline][Order article via Infotrieve]
- Flynn J, Slovic P, Mertz CK. Gender, race, and perception of environmental health risks. Risk Anal. 1994;14:1101–1108.[CrossRef][ISI][Medline][Order article via Infotrieve]
- Pelle E, Ingrassia M, Mammone T, et al. Protection against cigarette smoke-induced damage to intact transformed rabbit corneal cells by N-acetyl-L-cysteine. Cell Biol Toxicol. 1998;12:253–259.
- Basu PK, Pimm PE, Shephard RJ, Silverman F. The effect of cigarette smoke on the human tear film. Can J Ophthalmol. 1978;13:22–26.[Medline][Order article via Infotrieve]
- Sopori ML, Kozak W. Immunomodulatory effects of cigarette smoke. J Neuroimmunol. 1998;83:148–156.[CrossRef][ISI][Medline][Order article via Infotrieve]
- Begley CG, Riggle A, Tuel JA. Association of giant papillary conjunctivitis with seasonal allergies. Optom Vis Sci. 1990;67:192–195.[Medline][Order article via Infotrieve]
- Scheme European Influenza Surveillance. 2004;Available at http://www.eiss.org/. Accessed on November 25
Top
Abstract
Methods
