Risk factors for acanthamoeba keratitis in contact lens users: a case-control study

Introduction

Acanthamoeba keratitis is an uncommon but potentially devastating corneal infection. In a recent series, 34 corneal grafts were performed in 23 of 73 eyes (32%), and for 15 of the 72 patients (21%) the characteristically recalcitrant pain and protracted clinical course culminated in severe visual loss.1

Although acanthamoeba keratitis occurs after contaminated corneal injuries,2 85% of cases are in contact lens users of all lens types.3 Conventional soft lens wearers account for the majority of cases, reflecting the popularity of this lens type, in whose users the use of homemade saline, swimming in lenses, and irregular disinfection are risk factors.4

Acanthamoebas are free living amoebas found in air, dust, and water, including bathroom tap water.5 Their cysts are resistant to antimicrobial agents, including current lens disinfection solutions,6 and may contaminate lens storage cases in asymptomatic users.7

At Moorfields Eye Hospital there has been an increase in the number of acanthamoeba cases corresponding to the introduction and increasing popularity of disposable lenses.1 We carried out a retrospective case-control study to investigate causes for the increase in acanthamoeba keratitis cases in London. We determined the relative risks of acanthamoeba keratitis associated with different lens types and the relative importance of lens type, disinfection regimens, and other risk factors.

Subjects and methods

METHODS

Cases and controls completed a self administered questionnaire detailing occupation, age, sex, lens use, experience, and hygiene practices. These data were confirmed by telephone when they were insufficient or ambiguous. Socioeconomic classification was carried out using the 1991 census coding.10 Extended wear was defined as overnight use occurring at least once per week, although less frequent overnight use was included in the analysis. For each individual a hygiene score for each aspect of lens care was derived. Daily mechanical lens cleaning, disinfection with fresh solution on every lens removal, weekly case cleaning, and fortnightly enzyme treatment (or equivalent disposal frequency) scored maximally. Points were deducted for less frequent or suboptimal attention to these aspects of lens care. Use of non-sterile water for any aspect of lens care received a zero score for disinfection. For the statistical analysis each step in the hygiene regimen was classified dichotomously as optimal or suboptimal. Clinical data were derived from hospital notes and previous study data.1

STATISTICAL ANALYSIS

The EGRET software package (Statistics and Epidemiology Research Corporation, Seattle) was used to carry out multivariable logistic linear regression analysis. STATXACT (CYTEL Software Corporation, Cambridge, MA) was used for estimation of exact odds ratios (relative risks), 95% confidence intervals, and P values. The population attributable risk percentage, a measure of the proportion of cases that can be attributed to a specific risk factor, was estimated where appropriate.

Results

Thirty nine patients with lens related acanthamoeba keratitis were identified. Four were excluded: one therapeutic lens user and three patients, attending before 1 September 1991, who could not be traced for verification of lens wear data. Results of tissue cultures were positive for 14 patients (40%), and Acanthamoeba was isolated from the lens, lens case, or solutions from a further eight patients (23%). Table I shows the characteristics of the study population.

RISK WITH DIFFERENT LENS TYPES

Daily wear disposable soft lenses were used by 9/12 (75%) of the cases and 37/378 (9.8%) of the controls (table II). The crude relative risk associated with daily wear disposable as opposed to conventional daily wear soft lenses was estimated as 49.45 (95% confidence interval 6.53 to 2227, P<0.001).

HYGIENE PRACTICES AND OTHER RISK FACTORS

As the use of daily wear soft lenses was associated with 31 of the 35 cases, analysis of further risk factors was restricted to patients using daily wear conventional or disposable soft lenses. Five controls who used heat disinfection were excluded from this analysis because no cases used this method.

The 16 variables incorporated in the analysis of risk factors for this group are listed in the box. The multivariate model was constructed with a step up procedure, in which significant variables were added one at a time to assess their effect on the fit of the model.

The use of non-sterile saline or water could not be incorporated into the model as it was too closely linked with the disinfection score. Separate analysis, however, showed a significantly increased risk of acanthamoeba keratitis (12.51 (1.37 to 156), P=0.023) associated with its use. The risk of swimming in lenses, which occurred in three of the cases, was not evaluated as these data were not available for the controls.

The type of disposable lens used was also excluded from the main analysis, owing to insufficient numbers using types other than Acuvue (Vistakon, UK). Among patients wearing daily wear disposable soft lenses all 19 cases and 24/37 (65%) of the controls were using Acuvue. A crude minimum relative risk for Acuvue compared to other disposable lens types was estimated as 2.63 (2.63 to (infinity); P=0.004).

Table III shows the relative risk and 95% confidence intervals of the most important exposure factors. Although the crude analysis of risk with different lens types showed a 49.45 increased risk associated with daily wear disposable lenses, multivariate analysis showed that the excess risk attributable to the lenses is comparatively small (3.82, P=0.049).

The factor with the greatest effect on the risk of keratitis was the type and standard of disinfection used. As expected, failure to disinfect carried a greatly increased risk of keratitis (55.86, P<0.001), but chlorine based disinfection, even when used optimally, was also shown to carry a significantly increased risk compared with hydrogen peroxide or other chemical systems (14.63, P=0.001). The population attributable risk percentage, after lens type was controlled for, was 56.3% (27% to 86%) for chlorine based disinfection and 31.8% (8 to 55) when disinfection was omitted.

MULTIVARIATE ANALYSIS WITH CULTURE POSITIVE CASES

Because of the small number of cases with positive results of corneal culture, and because of the absence of cases among the referent group exposed to some factors, analysis was carried out by stratifying the data by a few important extraneous factors and using exact procedures for significance tests and estimations of relative risks. Table IV shows that exclusion of cases with negative results on tissue culture gives similar or slightly increased lower confidence limits for the excess risks associated with the use of chlorine disinfection and omitting disinfection although, with the smaller number of cases, the excess risk associated with disposable lens use was not statistically significant.

Discussion

We have shown that the use of daily wear disposable lenses is associated with a greatly increased risk of acanthamoeba keratitis. Multivariable analysis showed that this is largely attributable to a lack of disinfection, the use of non-sterile saline, and the use of chlorine based disinfection rather than alternative chemical systems. Case-control studies have also shown an increased risk for bacterial keratitis with disposable lenses.8 11 In one study this was attributed largely to overnight wear, rather than to the effects of lens hygiene and the lens itself11 that we have shown for acanthamoeba keratitis.

DISPOSABLE LENSES AND ACANTHAMOEBA KERATITIS

A relatively small excess risk was associated with disposable lenses independent of other risk factors. The reasons for this may include confounding factors that we could not assess, such as the use of disposable lenses as a “problem solver” in patients at higher risk of keratitis. Alternatively, physical factors such as increased in vivo lens dehydration12 and protein absorption,13 due to the high water content ionic material from which many of these lenses are made, or the presence of manufacturing defects affecting corneal integrity14 might account for this finding.

DISINFECTION SYSTEMS AND ACANTHAMOEBA KERATITIS

Chlorine has been compared with other disinfection solutions and shown to increase the risk for acanthamoeba keratitis; it accounted for 56% of cases. It is also less protective against bacterial keratitis.15 acanthamoeba cysts can survive 10 times the concentration of chlorine used in these systems,16 and chlorine based disinfection has been associated with contamination of storage cases by acanthamoeba.17 Other chemical systems are either ineffective against acanthamoeba cysts or effective only in the absence of organic debris or after a prolonged soaking time.6 However, since concomitant bacterial contamination may be important to the survival and growth of acanthamoeba in the lens case,18 the relative protection apparently provided by these systems may be due to their greater antibacterial efficacy.

Comparative in vitro studies on the efficacy of chlorine release systems against a panel of bacteria and fungi have shown relatively poor activity19 and equivalent activity.20 These differing results may be due to interstrain variability in susceptibility to disinfectants. Furthermore, these in vitro tests, used by licensing authorities for premarket evaluation of the efficacy of contact lens disinfectants in Britain,21 probably do not reflect the relative efficacy of these agents in vivo, where organisms contaminating both contact lenses and cases have been shown to live in biofilms.22 Organisms in biofilms exist as a sessile phenotype and are more resistant to disinfectants than the planktonic phenotype used in the laboratory for in vitro tests.23

Although acanthamoebas were isolated from the lens storage cases of half the patients with acanthamoeba keratitis, lens case hygiene did not affect the risk of developing keratitis. Given the misunderstandings regarding lens case hygiene,24 many people may have been using inappropriate methods that increased the risk of lens case contamination.5 Since this study was conducted, the manufacturers of Softab (Alcon Laboratories, UK), whose product accounted for 85% of the chlorine based disinfection in this study, have repackaged their system to include a new storage case with each month's supply of tablets. This may reduce both the prevalence of contamination of storage cases and the excess risk associated with the use of chlorine based disinfection.

PREVENTING ACANTHAMOEBA KERATITIS

Over 80% of cases of acanthamoeba keratitis could be prevented by the adequate use of an effective disinfection system. This avoidable burden of disease seems to have two major causes: the “low care” philosophy of daily disposable lens use has resulted in “no care” practised by ill informed contact lens practitioners and users, and the easy to use chlorine release systems, commonly dispensed with disposable lenses, seem to be particularly ineffective against Acanthamoeba. Although contact lens disinfection systems in Britain have to comply with stringent tests of efficacy against a panel of microbes, these are directed against planktonic organisms in vitro, rather than the more resistant sessile organisms that are found in vivo and do not include Acanthamoeba.21 Increased emphasis on the importance of maintaining good hygiene with soft conventional and reused disposable lenses, using hydrogen peroxide or other disinfection systems that are effective against Acanthamoeba, could almost eliminate this severe cause of keratitis. Contact lens practitioners, the contact lens industry, and the Department of Health all have roles to play in achieving this aim—not least by holding back licences from disinfection systems that are ineffective against Acanthamoeba.

We thank Mr Ronald Martin for clerical assistance, the nursing staff of Moorfields accident and emergency department for their help in conducting the study, and Drs Mark Elder and Eric Beck for reviewing the manuscript. This study was supported by a grant from Alcon Laboratories (UK) Ltd.