Content uploaded by Mark A Bullimore
Author content
All content in this area was uploaded by Mark A Bullimore on Dec 28, 2018
Content may be subject to copyright.
The Risk of Microbial Keratitis With Overnight
Corneal Reshaping Lenses
Mark A. Bullimore*, Loraine T. Sinnott
†
, and Lisa A. Jones-Jordan
‡
ABSTRACT
Purpose. To estimate the incidence of microbial keratitis (MK) associated with overnight corneal reshaping contact lenses
and to compare rates in children and adults.
Methods. A retrospective study of randomly selected practitioners, stratified by order volume and lens company, was
conducted. Practitioners were invited to participate and those agreeing were asked to provide deidentified patient infor-
mation for up to 50 lens orders and to complete a comprehensive event form for any of these patients who have attended an
unscheduled visit for a painful red eye. Duration of contact lens wear was calculated from the original fitting date or January
2005 (whichever was later) to when the patient was last seen by the practitioner wearing the lenses on a regular basis. Cases
of MK were classified by majority decision of a 5-member expert panel.
Results. For the 191 practitioners who could be contacted, 119 (62%) agreed to participate. Subsequently, 11 withdrew, 22
did not respond, and 86 (43%) returned completed forms corresponding to 2202 lens orders and 1494 patients. Limiting the
sample to those patients with at least 3 months of documented contact lens wear since 2005 resulted in a sample of 1317
patients; 640 adults (49%) and 677 children (51%) representing 2599 patient-years of wear (adults = 1164; children = 1435).
Eight events of corneal infiltrates associated with a painful red eye were reported (six in children and two in adults). Two
were classified as MK. Both occurred in children but neither resulted in a loss of visual acuity. The overall estimated in-
cidence of MK is 7.7 per 10,000 years of wear (95% confidence interval [CI] = 0.9 to 27.8). For children, the estimated
incidence of MK is 13.9 per 10,000 patient-years (95% CI = 1.7 to 50.4). For adults, the estimated incidence of MK is 0 per
10,000 patient-years (95% CI = 0 to 31.7).
Conclusions. The risk of MK with overnight corneal reshaping contact lenses is similar to that with other overnight mo-
dalities. The fact that the CIs for the rates estimated overlap should not be interpreted as evidence of no difference. True
differences fewer than 50 cases per 10,000 patient-years were beyond the study’s power of detection.
(Optom Vis Sci 2013;90:937Y944)
Key Words: microbial keratitis, overnight orthokeratology, contact lenses, extended wear, adverse effects
Corneal reshaping, also known as corneal refractive therapy
or orthokeratology (OK), was initially reported in the
early 1960s.
1
The original goal of corneal reshaping was to
permanently change the shape of the cornea while wearing contact
lenses for several hours early in the day. Early studies reported an
incomplete treatment effect and transient, unpredictable refractive
error reduction.
2
In the mid-1990s, innovative materials and reverse-geometry
designs coupled with videokeratography allowed for quicker,
more predictable treatment effects and nighttime contact lens
wear.
3
Corneal reshaping contact lenses are now typically worn
during sleep to temporarily reduce myopia by flattening of the
cornea. A summary of the history, safety, and effectiveness of corneal
reshaping is provided in Swarbrick’s
4
comprehensive review.
Reports of microbial keratitis (MK) associated with corneal
reshaping contact lens wear, particularly in children, have caused
concern about the safety of this modality. While these reports,
summarized by Watt and Swarbrick,
5
highlight the importance of
continued monitoring of complications associated with corneal
reshaping contact lens wear, they do not allow comparison of the
risk of severe complications associated with corneal reshaping
contact lenses to other contact lens modalities. Because a low
proportion of the population wears overnight corneal reshaping
contact lenses, large-scale studies usually identify no cases of MK
in this group.
6,7
1040-5488/13/9009-0937/0 VOL. 90, NO. 9, PP. 937Y944
OPTOMETRY AND VISION SCIENCE
Copyright *2013 American Academy of Optometry
ORIGINAL ARTICLE
Optometry and Vision Science, Vol. 90, No. 9, September 2013
*MCOptom, PhD, FAAO
†
PhD
‡
PhD, FAAO
College of Optometry (MAB), The University of Houston, Houston, Texas;
and College of Optometry (LTS, LAJ-J), The Ohio State University, Columbus,
Ohio.
Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.
Many of the cases of corneal ulcers that have been published
have reported data from children; however, it should not be as-
sumed that children have a greater risk of complications with
corneal reshaping contact lens wear. Complications in children
may be reported more often than in adults because of the greater
potential for adverse effects due to a larger number of cumulative
years that a young person may be exposed to risk. There may
also be more children wearing corneal reshaping contact lenses
than adults owing to the potential for myopia control with the
contact lenses.
8Y10
The risks of corneal reshaping contact lens wear
in children cannot be compared to the risks in adults using only
the data currently published in the literature. Large-scale, retro-
spective or prospective studies of overnight corneal reshaping
contact lens wear in general, specifically in children, have not as
yet been reported.
There are insufficient data on the absolute frequency of MK
in overnight OK or on the relative risks compared with other
contact lens modalities.
6
Nonetheless, editorials and opinion pieces
have questioned the safety of corneal shaping lenses, particularly
in children.
11Y13
In 2002, Paragon Corneal Refractive Therapy lenses, manu-
factured and distributed exclusively by Paragon Vision Sciences,
were granted Food and Drug Administration (FDA) approval, al-
though other lens designs and materials are covered by the original
approval. In 2004, Bausch + Lomb received approval for the
Boston Vision Shaping Treatment lens. These contact lenses are
marketed as the Contex E series, the Euclid Emerald Lens, the
DreamLens, and the BE Lens. In 2006, the FDA issued Section
522 orders to both companies mandating that they conduct
Postmarket Surveillance of their respective corneal reshaping lenses
to address concerns about the use of these lenses in children. Spe-
cifically, the orders required the companies to address ‘‘in patients
undergoing overnight OK treatment, what is the relative risk of
developing MK in persons under the age of 18 as compared to
adults?’’ Using a retrospective cohort of patients fitted with over-
night corneal reshaping lenses, this study compared the incidence
of MK in children and in adult patients.
METHODS
In 2005, the FDA Office of Surveillance on Biometrics
contacted Bausch + Lomb and Paragon requesting information on
the incidence of MK, with particular reference to children com-
pared to adult patients, and indicated their intention to require the
conduct of a postmarket surveillance study. All parties agreed that
a well-designed retrospective study would be able to address the
core area of concern to the FDAVthe risk of MK in patients
younger than 18 years, relative to adults. There was identification
of a number of areas of importance, including
&The need to determine the exposure time (duration of wear) of
each patient
&The need to ascertain lost to follow-up patient data
&Consideration of incentives for practitioners to contact patients
who have not returned to the practice and the collection of
meaningful data from these patients
&The selection of participating practices including randomiza-
tion and stratification by number of corneal reshaping patients
&The need for specific criteria for the diagnosis of MK
The protocol for the present study was developed by the authors
and approved by the FDA and by the Ohio State University (OSU)
Office of Research Risks and Protection. Informed consent was
obtained as described below. The study was conducted in 2007.
The goal of the study was to compare the incidence of MK in
children and in adult patients wearing corneal reshaping lenses
using a retrospective cohort of patients fitted with these lenses in
2005 and 2006. The incidence in each age group was estimated
based on the number of cases (numerator) and the years of lens
wear (denominator). Years of lens wear were the total years of wear
accumulated in the group. The difference in risk was estimated by
comparing the incidence in the two age groups.
The sample-size goals of the study were to identify, through 200
practitioners, 1000 randomly selected patients in each age group,
with sufficient follow-up to provide a total of 2000 patient-years
of exposure across the two groups. With 2000 patient-years, if the
true adult rate of incidence was 10 cases in 10,000 years and the
true rate in children was 60 cases in 10,000 years, a difference
could be detected with at least 80% power. These rates are similar
in magnitude to those for other overnight wear modalities.
6
Selection and Recruitment of Practitioners
Paragon provided OSU investigators with a comprehensive
database containing names of practitioner accounts. For each
account, they provided a history of all lens orders from 2005 and
2006. The information included order date and lens parameters.
Similar information was provided by five authorized lens finish-
ing laboratories licensed to sell corneal reshaping lenses under
Bausch + Lomb’s FDA Premarket Approval.
It is possible that the incidence of MK varies with practice
volume, in terms of number of corneal reshaping lenses fitted.
Therefore, for each company, a sampling strategy was developed,
which targeted both practitioners fitting a large number of pa-
tients with corneal reshaping lenses and those fitting relatively few
patients. The random sampling strategy recruited equal numbers
of low- and high-volume practitioners, but limited the number of
patients contributed by any one practitioner to 50, to minimize
the respondent burden and to avoid any single practice contrib-
uting a substantial proportion of the sample.
The Optometry Coordinating Center at OSU coordinated
all stratification with safeguards to mask both companies. The
following protocol applied to each company. First, a complete
list of practitioners along with their contact information and
the number of orders in 2005 and 2006 was obtained. The
practitioners were stratified into high- (Q25 orders) and low-
(G25 orders) volume groups. Within each stratum from each
volume group and each company, a systematic sample was chosen,
selecting every kth name, where kwas determined by the desired
sample size and the available population (k= population size/sample
size). Second, for all practitioners, complete lens order data from
2005 and 2006 were obtained from the company. If 25 or fewer
lens orders were available, all lens orders were sampled. For those
practitioners with at least 25 lens orders, up to 50 orders were
selected using simple random sampling methods. Lens order
data for all practitioners were obtained from both companies (and
affiliate lens finishing laboratories).
Paragon Vision Science mailed letters to a large group of
practitionersVequal numbers of high and low volumeVselected at
938 Microbial Keratitis Risk With Overnight Corneal Reshaping LensesYYBullimore et al.
Optometry and Vision Science, Vol. 90, No. 9, September 2013
Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.
random from their customer database. Of these, 100 were later
randomly selected for participation by the OSU investigators.
The letter contained a description of the study, an indication that they
may be approached by the OSU investigators, encouragement to
participate, and assurances about confidentiality of participation and
patient data. A similar letter was sent to practitioners by the fol-
lowing participating Bausch + Lomb laboratoriesVArt Optical,
Essilor, Euclid, Precision Technology Services, and TruformVof
whom 100 were later randomly selected for participation.
The 200 randomly selected practitioners were contacted by
telephone by the principal investigator (MAB). When the prac-
titioner was not available, a comprehensive message was left. At
least three attempts were made to contact the practitioners by
telephone. In addition, an e-mail was also sent to the practitioner
when an address was available.
Data Forms
Participating practitioners were sent a customized Practitioner
Lens Summary Form and asked to complete and return it to OSU.
The form listed up to 50 pairs of corneal reshaping lenses ordered
in 2005 and 2006. Practitioners were requested to provide the
fitting date (quarter and year) and patient age at fitting for each
of the pairs of lenses listed. In addition, they were asked whether
the patient continued to wear the lenses and when the patient was
last seen in the practice. Finally, the practitioner was asked to
state whether each patient had experienced an episode of possible
MK, defined as a painful red eye that required a visit to a doctor’s
office. For each case of possible MK, the practitioner completed
an Event Form and returned it to OSU. This was intended to result
in over-reporting so that no potential cases were missed. The prac-
titioner may have been aware of cases in patients before the study
period or in patients fitted during the study period, but not included
on the list of lens orders. Such cases were ineligible for data analysis.
If the patient was treated by another practitioner, the patient
was mailed a prepared packet by the practitioner containing
Health Insurance Portability and Accountability Act and consent
forms and a questionnaire regarding months of lens wear and any
experienced adverse events. The form requested the name and
address of the treating doctor. Patients were asked to return these
three forms to OSU investigators. If the practitioner already knew
the name of the treating doctor, the patient was not asked for the
information. Only the deidentified summary information re-
quested on the event form should have been provided, and in-
dividual patients were not contacted by OSU. The only identifiers
on the form were fitting date and patient age.
Patients with less than 12 months of documented follow-up
were mailed packets by the practitioner containing Health In-
surance Portability and Accountability Act and study consent
forms and a questionnaire regarding months of lens wear and any
experienced adverse events. Patients were asked to return these
three forms to OSU investigators in an accompanying prepaid
envelope. This strategy was used in an attempt to capture any cases
of discontinued wear because of MK.
Adjudication and Classification of Cases
An Outcomes Assessment Panel reviewed each case of possible
MK independently and determined by majority vote whether
the case was definite MK, probable MK, probably not MK,
definitely not MK, or MK unrelated to contact lens wear (see
below for details on classification of cases). The Outcomes As-
sessment Panel members were chosen based on their expertise in
the assessment, evaluation, and management of contact lens-
related complications. During classification of possible cases of
MK, the panel was masked to the age of the patient to remove the
potential for bias. The criteria for classification were similar to
those used in a previous study of the incidence of MK.
14
The
definition of definite MK was one or more corneal stromal in-
filtrates greater than 1 mm in size with pain more than mild and
one or more of the following:
&anterior chamber reaction more than minimal,
&mucopurulent discharge,
&positive corneal culture,
and treatment consistent with standard of care for MK in terms
of choice of medication, frequency, and duration.
The presence of a subsequent corneal scar was a requirement in
cases in which adequate follow-up records were available. In the
absence of positive or negative data concerning some of the firstthree
criteria, the level of treatment was heavily weighted in the Outcomes
Assessment Panel’s determination of the presence of MK.
The definition of probable MK was the same as that of definite
MK but failing to meet all of the specified criteria, for example,
the size of the lesion was less than 1 mm, the pain was minimal, or
there was no anterior chamber reaction, mucopurulent discharge,
or positive corneal culture.
The definitions of probably not MK and definitely not MK
were based on the extent that the above criteria were met. The
category of MK unrelated to contact lens wear was reserved for
cases, such as herpes simplex keratitis or staphylococcal marginal
keratitis, where, in the judgment of the Panel, the etiology was
unrelated to the wearing of contact lenses.
Data Analysis
The sample was described in terms of patient age, fitting date,
and duration of corneal reshaping contact lens wear. A patient
who was identified by multiple lens orders was only counted once
in the analyses.
The primary endpoints were as follows:
&the incidence of MK in children and adults and
&the difference in risk between the two age groups.
The incidence of MK in each age group was estimated as the
number of definite or probable MK cases divided by patient-years
of lens wear. The difference in risk was estimated by comparing
the incidence in the two patient age groups.
Only patients who had completed at least 3 months of wear
were included in the data analyses. Exposure was calculated based
on last patient contact with the practitioner. If less than 12 months
of follow-up were available, exposure was based on the last date
that the patient reported wearing the lenses. The incidence was
also estimated for all infiltrative events, regardless of severity.
Statistical analyses were performed using Statistical Analysis
Software (SAS, version 9.1.3) and StatXact (version 8). The pa-
rameter of central concern was the incidence rate, which is esti-
mated as the ratio of the number of incidents in a sample divided
by the total person-time accumulated by the sample. Estimates of
this parameter, as well as descriptive summaries (means and
Microbial Keratitis Risk With Overnight Corneal Reshaping LensesYYBullimore et al. 939
Optometry and Vision Science, Vol. 90, No. 9, September 2013
Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.
standard deviations) of practitioner and patient characteristics,
were computed in SAS. Exact confidence intervals (CIs) for the
absolute value of incidence rates were also computed in SAS using
the method of Garwood.
15
Exact CIs for the difference in the
incidence rates between children and adults were computed in
StatXact using the method of Chan and Zhang.
16
RESULTS
Of the 200 randomly selected practitioners, 9 could not be
contacted because of illness, practice closure, or because the
practitioner had left the practice. For the 191 practitioners who
could be located and contacted, 119 agreed to participate (62%).
For the 119 practitioners who agreed to participate, 66 were
high volume (55%) and 53 were low volume (45%). Subsequent
to agreeing to participate, 86 practitioners returned completed
forms, 11 withdrew, and 22 failed to complete the study forms
despite multiple and frequent reminders. There was no significant
difference (W
2
test, p = 0.085) between the proportion of high-
volume practitioners (51 of the 99 contacted) and the low-
volume practitioners submitting data (35 of the 92 contacted).
The completed forms corresponded to 2202 lens orders and
represented 1494 unique patients.
Practitioners sent 48 packets to patients who had not been
followed for at least 12 months. Of these, 22 patients returned the
completed forms to investigators, and 14 of these provided new
information regarding duration of lens wear.
Duration of Lens Wear
Patients were identified based on lens orders in 2005 and 2006
without differentiating among patients fitted for the first time in
an approved corneal reshaping lens, patients refit in a new design,
or lenses ordered for other reasons. Some patients had docu-
mented lens wear before 2005, but only data on lens wear from
2005 onward and cases of possible MK from 2005 onward were
analyzed to minimize bias.
For most patients, duration of lens wear was based on the date
the patient was last seen in the practitioner’s office wearing their
lenses on a regular basis. For patients submitting forms containing
follow-up information, the duration of lens wear reported by the
patient was used. In patients for whom the practitioner reported
an original fitting date in 2005 or after, the duration of lens wear
was based on this date. In patients for whom the practitioner
reported an original fitting date before 2005, the duration of lens
wear was calculated conservatively from January 2005.
The above criteria resulted in a sample of 1317 patients: 640
adults (49%) and 677 children (51%). These patients are dis-
tributed relatively evenly between the two companies (697 vs.
620), and represent a total of 2599 patient-years of wear: 1164 in
adults and 1435 in children. The patient-years are distributed
relatively evenly between the two companies (1374 vs. 1224).
Table 1 summarizes the number of participating practitioners
stratified by practice volume and company. The number of pa-
tients and contributed patient-years are also shown, stratified by
age group, practice volume, and company. At the original fitting
date, the mean (SD) age of the adults was 38.0 (11.1) years and the
mean (SD) age of the children was 12.2 (2.5) years. The age
distribution of the children and adults is shown in Fig. 1. The
mean (SD) follow-up for the adults was 1.8 (1.0) years, with 497
(78%) having at least 12 months of follow-up. The mean (SD)
follow-up for the children was 2.1 (0.8) years, with 620 (92%)
having at least 12 months of follow-up.
Incidence of MK
A total of 50 event forms were submitted. These were from
27 practitioners, representing almost one third of those providing
patient data. All forms were transcribed and sent to the five mem-
bers of the Outcome Assessment Panel for adjudication. Eleven
forms reported a corneal infiltrate of which seven were in children,
but as described previously, panel members were masked to the age
of the patient. Three of the cases of infiltrates occurred before 2005
and thus not within the years of lens wear used to determine risk. Of
the eight qualifying infiltrative events, six occurred in children.
The results of the Panel’s assessment are summarized in Table 2.
Two cases were classified as definite MK by the Panel and used for
the primary analyses. Thus two cases of MK occurred within 2599
patient-years of overnight corneal reshaping lens wear. Neither of
the MK cases resulted in any documented long-term loss of best-
corrected visual acuity. One adult was not examined again after
an event but was classified by the Panel as definitely not MK.
Table 3 summarizes the incidence estimates for MK based on
the 2599 patient-years. The overall estimated incidence of MK is
7.7 per 10,000 patient-years (95% CI = 0.9 to 27.8). For children,
the estimated incidence of MK is 13.9 per 10,000 patient-years
(95% CI = 1.7 to 50.4). For adults, the estimated incidence of
MK is 0 per 10,000 patient-years (95% CI = 0 to 31.7).
The difference in rates (children jadults) was also calculated
along with the CI. The 95% CI includes zero, so there is no sta-
tistically significant difference in the rates between children and
adults (difference = 13.9 per 10,000 patient-years, 95% CI = j17.9
to þ50.9).
A conservative analysis limited to the 685 patients (representing
1415 patient-years) who began wear of overnight corneal reshaping
lenses after January 2005 and had at least 12 months of docu-
mented lens wear is presented in Table 4. The overall estimated
incidence of MK is 14.1 per 10,000 patient-years (95% CI = 1.7
to 51.1). For children, the estimated incidence of MK is 25.2
per 10,000 patient-years (95% CI = 3.1 to 91.0). For adults, the
TABLE 1.
Summary of participating practitioners and eligible patients
High
volume
Low
volume
Company
A
Company
B Total
Number of
practitioners
51 35 47 39 86
Patients
contributed
1166 151 697 620 1317
Adults 554 86 383 257 640
Children 612 65 314 363 677
Patient-years
contributed
2329 270 1374 1224 2599
Adults 1017 147 711 452 1164
Children 1312 123 663 772 1435
940 Microbial Keratitis Risk With Overnight Corneal Reshaping LensesYYBullimore et al.
Optometry and Vision Science, Vol. 90, No. 9, September 2013
Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.
estimated incidence of MK is 0 per 10,000 patient-years (95%
CI = 0 to 59.4). For this more conservative criterion, there is
again no statistically significant difference in the rates between
children and adults (difference = 25.2 per 10,000 patient-years,
95% CI = j34.1 to þ92.7).
Incidence of Corneal Infiltrates Associated
With a Painful Red Eye
Table 5 summarizes the incidence estimates for all infiltrative
events based on the 2599 patient-years. The overall estimated
incidence of corneal infiltrates is 30.8 per 10,000 patient-years
(95% CI = 13.3 to 60.7). For children, the estimated incidence
of corneal infiltrates is 41.8 per 10,000 patient-years (95% CI =
15.3 to 91.0). For adults, the estimated incidence of corneal in-
filtrates is 17.3 per 10,000 patient-years (95% CI = 2.1 to 62.1).
The rate difference is not statistically significant (difference = 24.5
per 10,000 patient-years, 95% CI = j23.9 to +76.8).
DISCUSSION
Eight events of corneal infiltrates in patients presenting for an
unscheduled visit with a painful red eye were documented in 1,317
patients wearing overnight corneal reshaping lenses for a total of
2599 patient-years. Two events were judged to be MK by inde-
pendent masked experts. Both cases were in children and neither
resulted in a long-term loss of best-corrected visual acuity. The es-
timated overall incidence of MK is 7.7 per 10,000 patient-years
(95% CI = 0.9 to 27.8). The estimated incidence is 13.9 per
10,000 patient-years in children (95% CI = 1.7 to 50.4) and 0.0 per
10,000 patient-years in adults (95% CI = 0 to 31.7). The difference
in rates is 13.9 per 10,000 patient-years (95% CI = j17.9 to
þ50.9). In other words, the rate in children may be lower than the
rate in adults by up to 17.9 per 10,000 patient-years or the rate in
children may be higher than the rate in adults by up to 50.9 per
10,000 patient-years. The major findings of the study will be
summarized inthe package inserts and patient information booklets
for FDA-approved overnight corneal reshaping contact lenses.
The fact that the CIs for the rates estimated overlap should not
be interpreted as evidence of no difference. True differences under
50 cases per 10,000 patient-years were beyond the study’s power
of detection. Nonetheless, there is some evidence that the rate may
be higher in children. Six of the eight cases of corneal infiltrates
accompanied by a painful red eye occurred in children, and both
of the cases classified as MK were in children. The estimated
incidence of MK in children wearing overnight corneal reshaping
contact lenses is similar in magnitude to that previously associated
FIGURE 1.
Age distribution of patients at fitting.
TABLE 2.
Summary of classification of cases by the outcomes assessment panel
Original fit
(quarter/year)
Time from
original fit to event, mo Age at fitting, y Age at event, y BCVA at event BCVA at resolution Classification
3/2006 15 11 12 20/200 20/20 Definite MK
3/2005 15 15 17 20/20 20/20 Definitely not MK
3/2005 15 15 16 20/20 20/20 Definite MK
3/2005 5 9 9 20/20 20/20 Definitely not MK
1/2005 5 12 12 20/20 20/20 Probably not MK
3/2004 6 8 8 20/30 20/20 Definitely not MK
4/2005 10 27 28 20/30 20/20 Definitely not MK
4/2004 5 28 29 20/40 Lost to follow-up Definitely not MK
Note that although two patients were originally fit before 2005, the cases of corneal infiltrates occurred in 2005.
Microbial Keratitis Risk With Overnight Corneal Reshaping LensesYYBullimore et al. 941
Optometry and Vision Science, Vol. 90, No. 9, September 2013
Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.
with the overnight wear of soft contact lenses. Practitioners and
particularly parents should be aware of this risk because it is an
important part of the risk-benefit ratio.
Most previously published articles on MK associated with
overnight corneal reshaping lenses are case reports and small case
series.
5,17
In the largest previous study, Lipson
18
retrospectively
evaluated outcomes of overnight corneal reshaping in children and
adults. Data on 296 patients were reported, of whom 52% were
12 years or younger, representing 507 patient-years. He reported
three adverse events during the study. Although the definition of
adverse event included ‘‘microbial keratitis, a corneal abrasion
requiring medical treatment, loss of best-corrected visual acuity,
or a corneal scar,’’ none were MK (Lipson MJ, personal com-
munication, 2012). All three were presentations of superficial
punctate keratitis or epithelial defects greater than grade 2, which
did not result in a loss of best-corrected visual acuity, and all
three patients were still wearing their lenses at the conclusion of
the study. The two FDA premarket approval applications sub-
mitted by Paragon and Bausch + Lomb for approval of overnight
corneal reshaping lenses reported no cases of MK, although each
study reported one eye developing corneal infiltrates.
19,20
Col-
lectively, these studies enrolled 396 patients followed for up to
9 months, with around 70% completing 9 months of wear.
The present study is the largest attempt by far to quantify the
risk of MK associated with overnight corneal reshaping lenses with
2599 patient-years of wear. The above studies are smaller by
around a factor of 10. Given the relatively low incidence of MK,
it is not surprising that no cases were reported in these earlier
studies. Assuming an incidence of 7.7 per 10,000, it would re-
quire a study of around 1000 patient-years for the probability of
observing at least 1 case to exceed 50% (1 j(1 j0.00077)
1000
=
0.54) and more than 2000 patient-years for the probability of ob-
serving at least 1 case to exceed 80% (1 j(1 j0.00077)
2000
=0.80).
In other words, the two FDA premarket approval studies and that of
Lipson were unlikely to identify cases of MK based on the incidence
estimated in the present study.
Comparison With MK Rates in Other
Contact Lens Modalities
The incidence of MK associated with overnight corneal reshaping
lenses estimated in the present study is higher than that for daily
rigid contact lens wear but similar to that for overnight wear of soft
contact lenses, although corneal reshaping lenses are not worn for
most of the waking hours.
6,14,21Y24
A recent large-scale prospective,
12-month, population-based study estimated the risk of contact
lens-related MK.
6
The authors identified 285 eligible cases of
contact lens-related MK and 1798 controls. For daily wear of rigid
gas-permeable contact lenses, the annualized incidence was 1.2 per
10,000 wearers (95% CI = 1.1 to 1.5). Consistent with earlier re-
ports, the incidence for overnight wear of soft contact lenses was
higher: 19.5 per 10,000 wearers (95% CI = 14.6 to 29.5) for
conventional hydrogels and 25.4 per 10,000 wearers (95% CI =
21.2 to 31.5) for silicone hydrogels. Risk factors included overnight
use, poor storage case hygiene, smoking, Internet purchase of
contact lenses, less than 6 months of wear experience, and higher
socioeconomic status. The authors conclude that overnight use of
any contact lens is associated with a higher risk than daily use. The
authors of this large prospective population-based surveillance study
of contact lens-related corneal infection did not identify any cases of
MK associated with overnight corneal reshaping lenses (Stapelton, F
personal communication, 2009).
Study Limitations and Strengths
A number of factors should be considered when interpreting the
results of the present study. This was a retrospective study rather
than a prospective trial and thus prone to a number of limitations
TABLE 3.
Incidence of MK based on at least 3 months’ wear
Children Adults Overall
n 677 640 1317
Cases 2 0 2
Years at risk 1435 1164 2599
Incidence rate (95% CI) 0.00139 (0.00017 to 0.00504) 0 (0 to 0.00317) 0.00077 (0.00009 to 0.00278)
Rescaled incidence rate (95% CI) 13.9 (1.7 to 50.4) 0 (0 to 31.7) 7.7 (0.9 to 27.8)
Rate is per year of wear. Rescaled rate is per 10,000 patient-years.
CI, confidence interval.
TABLE 4.
Incidence of MK based on patients fitted after January 2005 and with at least 12 months’ wear
Children Adults Overall
n 378 307 685
Cases 2 0 2
Years at risk 794 621 1415
Incidence rate (95% CI) 0.00252 (0.00031 to 0.00910) 0 (0 to 0.00594) 0.00141 (0.00017 to 0.00511)
Rescaled incidence rate (95% CI) 25.2 (3.1 to 91.0) 0 (0 to 59.4) 14.1 (1.7 to 51.1)
Rate is per year of wear. Rescaled rate is per 10,000 patient-years.
CI, confidence interval.
942 Microbial Keratitis Risk With Overnight Corneal Reshaping LensesYYBullimore et al.
Optometry and Vision Science, Vol. 90, No. 9, September 2013
Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.
including, but not limited to, site and patient selection and par-
ticipation bias, and assumptions that disease incidence is constant
over time among wearers. Study sites were selected at random, but
participation was voluntary and practitioners who had observed
cases of MK may have been less willing to participate. The goal of
the study was to sample patients from 200 practitioners, but only
86 (43% of the goal) returned at least one completed form. In
retrospect, the protocol could have planned on replacing non-
participating practitioners using a predefined scheme based on
volume. The classification of MK was determined by an Outcome
Assessment Panel without direct contact with the patient or
photographs. Finally, the derived CIs are broad. The study’s
failure to find a statistically significant difference in the rate of MK
between adults and children may be due to a sample size too small
to detect a true underlying difference. As indicated earlier, a study
with 2000 patients and patient-years has an 80% chance to detect
a rate difference of 50 cases in 10,000 patient-years, a rate dif-
ference considerably larger than the 13.9 cases in 10,000 patient-
years observed in this study.
Any difference in the estimated rate of MK between children
and adults could have been biased by differential loss to follow-up
between the two age groups. Among the 1494 total patients, 677
of the 725 children (93%) had at least 3 months of follow-up
compared with 640 of 756 (85%) adults (W
2
= 28.6, pG0.001).
Among the 1317 patients with at least 3 months of documented
lens wear, 620 of the 677 children (92%) had at least 12 months of
follow-up compared with 487 of 640 (76%) of adults (W
2
= 58.9,
pG0.001). Thus, for both comparisons, follow-up was better in
children than in adults.
While previous published studies have reported individual or a
small series of overnight corneal shaping patients presenting with
MK, none have estimated its incidence. Despite the above limi-
tations, this was a large study (2599 patient-years) of randomly
selected patients fitted by randomly selected practitioners. Cases
were classified by an independent group of experts using clear and
established criteria and without knowledge of whether the patient
was a child or an adult.
Finally, it is important to emphasize that this study was
conducted in the United States, where clinical standards are high,
professional competencies are well maintained, and identification
and early treatment of adverse events may limit the impact of any
corneal infection.
Participation Rate and Sampling Strategy
The original sampling strategy of the study protocol was to
identify 100 randomly selected practitioners per company, and
1000 randomly selected patients in each age group with sufficient
follow-up to provide 2000 patient-years of exposure. Only around
60% of practitioners agreed to participate, with most of the
remaining practitioners not responding rather than declining to
participate. This could be a source of bias, although those prac-
titioners who did not participate did not differ from those who
did based on patient volume. The study relied almost entirely
on practitioners reporting corneal infiltrative events and a prac-
titioner may have been motivated not to report. Furthermore,
he/she may have been unaware of a patient who suffered an infec-
tion, was treated elsewhere and never returned to his practice. In
other studies of contact lens wear, extensive efforts have gone into
establishing a reliable numerator, that is, identifying all cases, for
example, by contacting the exposed subjects directly to ascertain if
they had experienced a painful, red eye. In this retrospective study,
this was not possible because patients had not given their informed
consent at the time of fitting to be contacted by a third party.
The goal of the sampling strategy was to obtain 2000 patient-
years of follow-up on 2,000 patients. The response of practitioners
resulted in 2599 patient-years but only 1317 patients. We con-
sidered sampling more practitioners or requesting data on more
patients from participating practitioners to reach 2000 patients.
However, either strategy could introduce more bias, without
providing substantially more statistical power. For example, the
majority of the practitioners (including all low volume practi-
tioners) reported data for all of their lens orders. Only 20% of
respondents (the highest volume practitioners) could provide
additional data from patients fitted during these years. Thus,
soliciting more patient data from respondents would have biased
the sample toward high-volume practitioners.
In planning this study, no data were available regarding the age
distribution of the patients fitted with overnight corneal reshaping
lenses, but the resulting sample demonstrates that children ac-
count for around half of the patients fitted with overnight corneal
reshaping lenses in the United States. Efron et al
25
reported a
survey of 105,734 contact lens fits in 38 countries from 2005 to
2009 including 13,926 in minors (G18 years of age). Overall,
corneal reshaping lenses accounted for 28% of the rigid lens fits in
minors, attesting to the popularity of this mode of correction for
younger age groups around the world. For 212,000 total contact
lens fits tracked by these authors since 1996, 1025 were described
as ‘‘rigid orthokeratology,’’ of whom 51% of patients are 17 years
and younger (Morgan PB, personal communication, 2012). In
other words, the distribution of overnight corneal reshaping lenses
between adults and children in the present study seems consis-
tent with practice worldwide. The viability of these lenses for the
TABLE 5.
Incidence of corneal infiltrates associated with a painful red eye based on at least 3 months’ wear
Children Adults Overall
n 677 640 1317
Cases 6 2 8
Years at risk 1435 1164 2599
Incidence rate (95% CI) 0.00418 (0.00153 to 0.0091) 0.00173 (0.00021 to 0.00621) 0.00308 (0.00133 to 0.00607)
Rescaled incidence rate (95% CI) 41.8 (15.3 to 91.0) 17.3 (2.1 to 62.1) 30.8 (13.3 to 60.7)
Rate is per year of wear. Rescaled rate is per 10,000 patient-years.
CI, confidence interval.
Microbial Keratitis Risk With Overnight Corneal Reshaping LensesYYBullimore et al. 943
Optometry and Vision Science, Vol. 90, No. 9, September 2013
Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.
control of myopia progression may have increased this proportion
in recent years.
8Y10
The American Academy of Ophthalmology published an
Ophthalmic Technology Assessment on overnight OK in 2008.
26
The authors conducted a thorough review of the literature and
identified 38 articles describing case reports or case series of MK
associated with overnight corneal reshaping contact lenses. The
authors acknowledge that the ‘‘value of the data is limited, be-
cause no denominator exists for the treatment population and the
total number of patients undergoing overnight orthokeratology is
unknown.’’ The authors state that ‘‘the prevalence and incidence
of complications associated with overnight orthokeratology have
not been determined’’ but that ‘‘the large number of children and
adolescents in these series and the risk of sight-threatening com-
plications in children and adolescents necessitates the highest level
of vigilance.’’ They conclude, ‘‘future research should be directed
at assessing the rate of infectious keratitis among overnight
orthokeratology users and whether the rate varies by age.’’ This
study is a first step in addressing this goal.
ACKNOWLEDGMENTS
The authors thank the 86 practitioners for their participation along with
the Outcomes Assessment Panel: Robin Chalmers, OD, Lisa Keay, BOptom,
PhD, Thomas Mauger, MD, Eric Ritchey OD, PhD, and Karla Zadnik,
OD, PhD. This study was funded by grants from Bausch + Lomb and
Paragon Vision Sciences to the Ohio State University. A number of in-
dividuals from these companies contributed to the design of the study, eval-
uation of the results and interactions with the FDA, including Kirk Bateman,
MS, Glenn Davies, OD, William Meyers, PhD, and William Reindel, OD,
MS. Mark A. Bullimore, MCOptom, PhD, is a paid consultant for Alcon
Research Ltd. and Carl Zeiss Meditec, Inc. Portions of this work were
presented at the 2009 Annual Meeting of the American Academy of Op-
tometry, Orlando, FL.
Received: December 4, 2012; accepted April 24, 2013.
REFERENCES
1. Jessen GN. Orthofocus techniques. Contacto 1962;6:200Y4.
2. Polse KA, Brand RJ, Schwalbe JS, Vastine DW, Keener RJ. The
Berkeley Orthokeratology Study, Part II: efficacy and duration. Am
J Optom Physiol Opt 1983;60:187Y98.
3. Nichols JJ, Marsich MM, Nguyen M, Barr JT, Bullimore MA.
Overnight orthokeratology. Optom Vis Sci 2000;77:252Y9.
4. Swarbrick HA. Orthokeratology review and update. Clin Exp
Optom 2006;89:124Y43.
5. Watt KG, Swarbrick HA. Trends in microbial keratitis associated
with orthokeratology. Eye Contact Lens 2007;33:373Y7.
6. Stapleton F, Keay L, Edwards K, Naduvilath T, Dart JK, Brian G,
Holden BA. The incidence of contact lens-related microbial keratitis
in Australia. Ophthalmology 2008;115:1655Y62.
7. Dart JK, Radford CF, Minassian D, Verma S, Stapleton F. Risk
factors for microbial keratitis with contemporary contact lenses: a
case-control study. Ophthalmology 2008;115:1647Y54.
8. Cho P, Cheung SW, Edwards M. The longitudinal orthokera-
tology research in children (LORIC) in Hong Kong: a pilot study
on refractive changes and myopic control. Curr Eye Res 2005;
30:71Y80.
9. Walline JJ, Jones LA, Sinnott LT. Corneal reshaping and myopia
progression. Br J Ophthalmol 2009;93:1181Y5.
10. Cho P, Cheung SW. Retardation of Myopia in Orthokera-
tology (ROMIO) study: a 2-year randomized clinical trial. Invest
Ophthalmol Vis Sci 2012;53:7077Y85.
11. Kwok LS, Pierscionek BK, Bullimore M, Swarbrick HA, Mountford
J, Sutton G. Orthokeratology for myopic children: wolf in sheep’s
clothing? Clin Experiment Ophthalmol 2005;33:343Y7.
12. Saviola JF. The current FDA view on overnight orthokeratology:
how we got here and where we are going. Cornea 2005;24:770Y1.
13. Schein OD. Microbial keratitis associated with overnight ortho-
keratology: what we need to know. Cornea 2005;24:767Y9.
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;112:2172Y9.
15. Garwood F. Fiducial limits for the poisson distribution. Biometrika
1936;46:441Y53.
16. Chan IS, Zhang Z. Test-based exact confidence intervals for the
difference of two binomial proportions. Biometrics 1999;55:1202Y9.
17. Santodomingo-Rubido J, Villa-Collar C, Gilmartin B, Gutierrez-
Ortega R. Orthokeratology vs. spectacles: adverse events and dis-
continuations. Optom Vis Sci 2012;89:1133Y9.
18. Lipson MJ. Long-term clinical outcomes for overnight corneal
reshaping in children and adults. Eye Contact Lens 2008;34:94Y9.
19. U. S. Food and Drug Administration Center for Devices and Radio-
logical Health. Paragon Vision Sciences. Summary of safety and effec-
tiveness data from Premarket Approval Application (PMA) supplement
number P870024/S43; 2002. Available at: http://www.accessdata.
fda.gov/cdrh_docs/pdf/P870024S043b.pdf. Accessed April 29, 2013.
20. U. S. Food and Drug Administration Center for Devices and Radio-
logical Health. Euclid Systems Corporation. Summary of safety and
effectiveness data from Premarket Approval Application (PMA) sup-
plement number P010062: 2004. Available at: http://www.accessdata.
fda.gov/cdrh_docs/pdf/P010062b.pdf. Accessed April 29, 2013.
21. 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;321:779Y83.
22. 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;89:430Y6.
23. Cheng KH, Leung SL, Hoekman HW, Beekhuis WH, Mulder PG,
Geerards AJ, Kijlstra A. Incidence of contact-lens-associated mi-
crobial keratitis and its related morbidity. Lancet 1999;354:181Y5.
24. 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. CLAO J 1994;20:
225Y30.
25. Efron N, Morgan PB, Woods CA. International survey of contact
lens prescribing for extended wear. Optom Vis Sci 2012;89:122Y9.
26. Van Meter WS, Musch DC, Jacobs DS, Kaufman SC, Reinhart WJ,
Udell IJ. Safety of overnight orthokeratology for myopia: a report by
the American Academy of Ophthalmology. Ophthalmology 2008;
115:2301Y13.
Mark A. Bullimore
356 Ridgeview Ln
Boulder, CO 80302
e-mail: bullers2020@gmail.com
944 Microbial Keratitis Risk With Overnight Corneal Reshaping LensesYYBullimore et al.
Optometry and Vision Science, Vol. 90, No. 9, September 2013
Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.
