Comparison of Latanoprost and Timolol in

Pediatric Glaucoma: A Phase 3, 12-Week,

Randomized, Double-Masked Multicenter
Study
Tomoko Maeda-Chubachi, MD, PhD,1 Katherine Chi-Burris, MPH,1 Brad D. Simons, MD, PhD,2
Sharon F. Freedman, MD,3 Peng T. Khaw, MD, PhD,4 Barbara Wirostko, MD,5 Eric Yan, PhD,6 for the
A6111137 Study Group*

Objective: To compare the efficacy and safety of latanoprost versus timolol in pediatric patients with
glaucoma.
Design: Prospective, randomized, double-masked, 12-week, multicenter study.
Participants: Individuals aged ⱕ18 years with glaucoma.
Methods: Stratified by age, diagnosis, and intraocular pressure (IOP) level, subjects were randomized (1:1) to
latanoprost vehicle at 8 AM and latanoprost 0.005% at 8 PM or timolol 0.5% (0.25% for those aged ⬍3 years) twice daily
(8 AM, 8 PM). At baseline and weeks 1, 4, and 12, IOP and ocular safety were assessed and adverse events were
recorded. Therapy was switched to open-label latanoprost PM and timolol AM and PM for uncontrolled IOP.
Main Outcome Measures: Mean IOP reduction from baseline to week 12. Latanoprost was considered
noninferior to timolol if the lower limit of the 95% confidence interval (CI) of the difference was ⬎⫺3 mmHg. A
proportion of responders (subjects with ⱖ15% IOP reduction at weeks 4 and 12) were evaluated. Analyses were
performed in diagnosis subgroups: primary congenital glaucoma (PCG) and non-PCG.
Results: In total, 137 subjects were treated (safety population; 12–18 years, n⫽48; 3–⬍12 years, n⫽55;
0 –⬍3 years, n⫽34). Mean age was 8.8⫾5.5 years, and mean baseline IOP was 27.7⫾6.17 mmHg; 125 subjects
completed the study, and 107 subjects were in the per protocol population. Mean IOP reductions for latanoprost
and timolol at week 12 were 7.2 and 5.7 mmHg, respectively, with a difference of 1.5 mmHg (95% CI, ⫺0.8 to
3.7; P⫽0.21). Responder rates were 60% for latanoprost and 52% for timolol (P⫽0.33). Between-treatment
differences in mean IOP reduction for PCG and non-PCG subgroups were 0.6 mmHg (95% CI, ⫺2.3 to 3.4) and
2.6 mmHg (95% CI, ⫺0.8 to 6.1), respectively. Responder rates for latanoprost versus timolol were 50% versus
46% for the PCG group and 72% versus 57% for the non-PCG group. Both therapies were well tolerated.
Conclusions: Latanoprost 0.005% is not inferior (i.e., is either more or similarly effective) to timolol and
produces clinically relevant IOP reductions across pediatric patients with and without PCG. Both latanoprost and
timolol had favorable safety profiles over the duration of this 3-month trial.
Financial Disclosure(s): Proprietary or commercial disclosure may be found after the references.
Ophthalmology 2011;xx:xxx © 2011 by the American Academy of Ophthalmology.

*Group members listed online in Appendix 1 (available at http://aaojournal.org).

Pediatric glaucoma, which encompasses a collection of di- that may occur at any age or be present at birth (e.g.,
verse pathophysiologic entities, is characterized by the pres- Sturge–Weber syndrome).1,3
ence of an abnormally elevated intraocular pressure (IOP) Depending on the cause and onset of pediatric glaucoma,
that represents a risk factor for optic nerve damage and the disease course, initial treatment options, subsequent
subsequent visual field loss. Classification of the childhood treatment, and follow-up can vary. Although children with
glaucomas into primary and secondary types provides a primary congenital glaucoma (PCG) usually undergo sur-
framework for the evaluation and management of individual gery soon after diagnosis,3–5 those with juvenile open-angle
patients.1,2 Primary glaucomas are those with isolated angle glaucoma (JOAG) and certain secondary glaucomas such as
malformations (trabeculodysgenesis) that have onset at birth aphakic glaucoma usually receive initial pharmacologic
(congenital), in the first few years of life (primary infantile), treatment.1,3
or later in childhood (juvenile). Secondary glaucomas are Reduction of IOP is the only proven and accepted
either acquired (e.g., aphakic glaucoma) or related to a method of slowing disease progression, and optimal IOP
clearly defined underlying ocular or systemic abnormality control is especially important given the long life expec-

© 2011 by the American Academy of Ophthalmology ISSN 0161-6420/11/$–see front matter 1

Published by Elsevier Inc. doi:10.1016/j.ophtha.2011.03.010
 Ophthalmology Volume xx, Number x, Month 2011

tancy (and therefore visual lifetime) of a pediatric patient adult dose of latanoprost ophthalmic solution (0.005%
with glaucoma. The last 30 years have seen the introduction dosed once daily) when compared with timolol in pediatric
of many effective agents for IOP management in adults, patients with glaucoma.
including nonselective and selective ␤-blocking agents,
prostaglandin analogs, topical and oral carbonic anhydrase
inhibitor preparations, and ␣-agonists.6,7 Few of these med- Materials and Methods
ications are approved for use in children. However, be-
cause of the clinical need to control IOP, medical treat- Study Design
ment of pediatric glaucoma often includes use of these
This was a 12-week, double-masked, parallel-group multicenter
topical and oral IOP-lowering medications alone or fre-
study (NCT00716859, EudraCT2007-004543-30). The final pro-
quently in combination.8,9 tocol and informed consent documentation were reviewed and
Latanoprost, a prostaglandin F2␣ analog, was the first of approved by the institutional review boards or independent ethics
its class to be approved for the reduction of elevated IOP committees at each participating investigational center. The study
associated with open-angle glaucoma or ocular hypertension. was conducted in compliance with the Declaration of Helsinki,
Latanoprost has become the most commonly prescribed first- with all International Conference on Harmonisation Good Clinical
line glaucoma therapy in adults.10 The prostaglandin agonists Practice guidelines, and, for sites in the United States, with the
and timolol are the most effective IOP-lowering agents when Health Insurance Portability and Accountability Act. Each sub-
meta-analysis is carried out.11 The long-term safety and effi- ject’s parent or legal guardian provided written informed consent
cacy of latanoprost in adult patients with elevated IOP are before study enrollment.
supported by more than 13 years of clinical experience, in-
cluding two 5-year safety studies.12,13 Although latanoprost is Subjects
not approved for use in children, there have been several
Eligible subjects were aged ⱕ18 years, diagnosed with glaucoma
publications that report the safety and effectiveness of its use in in 1 or both eyes, and had morning (before midday) IOP ⱖ22
pediatric patients with glaucoma.14 –19 A prospective, com- mmHg in at least 1 eye at the baseline visit. The agreement
passionate use study of 57 eyes of 48 children, of which 31 between the study sponsor and the European Medicines Agency
eyes had evaluable IOP data, showed that 32% of eyes had was to investigate the condition in the pediatric population, which
an IOP reduction of more than 10% after the addition of is defined as newborns, infants, infants and toddlers, children, and
latanoprost and only 19% had a reduction of more than adolescents. International Conference on Harmonisation/Commit-
15%.15 Responders were significantly more likely to have tee for Medicinal Products for Human Use guideline E11 specifies
JOAG.15 In a recently published retrospective chart review 5 age categories for possible evaluation within a pediatric inves-
by the same group,14 which included some of the same 48 tigational plan: preterm newborn infants; term newborn infants
children enrolled in the smaller prospective study, 115 chil- (0 –27 days); infants and toddlers (28 days to 23 months); children
(2–11 years); and adolescents (12 to 16 –18 years, depending on
dren had total latanoprost exposure of 2325 child-months. Of region). Preterm (⬍37 weeks gestation) newborn infants were
these 115 children, 63 eyes (63 children) had interpretable excluded from the study.
IOP data; 35% had an IOP reduction of ⱖ15% (mean Pharmacologic intervention with IOP-reducing therapy was
reduction, 27.7⫾8.4).14 Children with JOAG (n⫽19) had deemed clinically necessary in at least 1 eye. Prior glaucoma-
the highest response (74%) followed by those with aphakic related surgery, including drainage implants, was allowed as long
(n⫽13) glaucoma (46%), whereas those with PCG (n⫽14) as medical therapy was thought to be clinically indicated in this
or glaucoma associated with other diagnoses had a poor situation. Exclusion criteria are summarized in Table 1 (available
response. Systemic and ocular side effects associated with at http://aaojournal.org).
adult-dose latanoprost were infrequent and mild.14 To ensure a sufficient safety margin for systemic adverse
Over the last few years, there has been an increased focus effects after instillation of latanoprost 0.005% ophthalmic solution
once per day to the eyes in this pediatric population, enrollment
by regulatory agencies and pediatric advocacy groups on the was staged by age groups (12–18, 3–⬍12, and 0 –⬍3 years) after
need for the development and accessibility of medicinal examination of the results of the preceding pharmacokinetic study
products for use in the pediatric population. In the European (Invest Ophthamol Vis Sci, 51:E-abstract 210, 2010). In addition,
Union (EU), new pediatric legislation was enforced in 2007 a Data Monitoring Committee conducted an interim safety analysis
(Regulation [EC] No. 1901/200620) to improve the health of and could have recommended stopping the trial if safety concerns
children. In the United States, the Pediatric Research Equity were identified.
Act (Public Law No. 108-15521) was signed into law in
2003 to address the lack of pediatric drug use information in Treatments and Assessments
product labeling. These legislations were designed to facil-
itate timely pediatric medicinal product development and Potentially eligible subjects were evaluated at a screening visit that
the generation of data on the appropriate use of agents took place 0 to 28 days before the baseline visit. At screening,
approved for the treatment of adult diseases in pediatric ocular and medical histories were documented, a physical exami-
nation and biomicroscopy were performed, pupil diameter was
patients. To date, there have been no well-controlled large measured, and gonioscopy was performed if not done within the
trials to prospectively assess the safety and efficacy of preceding 6 months. In subjects able to cooperate, visual acuity
latanoprost or any other prostaglandin analog in the pediat- was measured and a visual field examination was conducted if
ric glaucoma population. Thus, the objective of this pro- clinically possible and if not performed within the prior 6 months.
spective, randomized, controlled trial was to establish, over For visual acuity measurement, children who were familiar with
3 months exposure, the relative safety and efficacy of the the letters of the alphabet were evaluated using Snellen visual

2
 Maeda-Chubachi et al 䡠 Latanoprost vs Timolol in Pediatric Glaucomas

acuity, and those who were unable or unfamiliar with the letters of ing the entire study period, measured the IOP using the same
the alphabet were evaluated using charts made up of numbers, tonometer at each visit for a given subject.
pictures, E’s, Landolt’s broken rings, and other methods that were At week 1 (and weeks 2 and 5, if needed), changes in medi-
equivalent to Snellen acuity. Intraocular pressure was measured in cations were recorded, resting pulse and blood pressure were
both eyes, with the right eye first, using Goldmann applanation measured, alertness was documented, hyperemia and visual acuity
tonometry (preferred method), Perkins tonometry, or a TonoPen were assessed, biomicroscopy was performed, and IOP was mea-
(Reichert, Inc., Depew, NY); the procedure was repeated for each sured. At week 4, week 1 assessments were repeated, pupil and
eye twice consecutively and a third time if the initial readings were corneal diameters were measured, and ophthalmoscopy was per-
inconsistent. Intraocular pressure was measured after visual acuity formed. At weeks 12 and 13 or at the end of treatment, all baseline
followed by biomicroscopy and finally pupil dilation. Subject assessments were repeated, visual field was evaluated, and gonios-
cooperation and any use of sedation during IOP measurements copy was performed.
were recorded. Investigators monitored all observed or volunteered adverse
After the screening visit, non-naive subjects (children who events throughout the study. Serious adverse events were those
were already receiving glaucoma medication) discontinued topical that were life-threatening; required inpatient hospitalization or
ocular hypotensive medications for at least 24 hours before the prolonged hospitalization; caused persistent or significant disabil-
baseline visit with longer washout periods at the discretion of the ity or incapacity; or resulted in congenital anomaly, birth defect, or
investigator. These subjects returned for IOP safety checks before death. Adverse events were followed until they resolved or stabi-
the baseline visit at the investigator’s discretion. The screening lized, and investigators assessed causality for each such event.
visit could be used as the baseline visit for those naive to ocular Masking codes were broken in emergency situations for reasons of
hypotensive medications. subject safety, and the investigator was to contact Pfizer before
At the baseline visit, any changes in medications were re- breaking the mask. If the masking code was broken, the reason was
corded; resting pulse and blood pressure were measured; alertness fully documented.
was documented using a 5-point scale (5 ⫽ responds readily to
name spoken in normal tone; 4 ⫽ lethargic response to name Statistical Analyses
spoken in normal tone; 3 ⫽ responds only after name is called
loudly or repeatedly; 2 ⫽ responds only after mild prodding and The primary efficacy end point was the mean IOP change from
shaking; and 1 ⫽ does not respond to mild prodding and shaking); baseline to week 12 in the study eye. At each visit, the median of
refraction (cycloplegic where appropriate) was performed; hyper- IOP measurements for the study eye represented the subject’s IOP.
emia was assessed by slit-lamp examination and graded (0 ⫽ none, For subjects who switched to open-label concomitant therapy or
who discontinued before week 12, the IOP measurement made at
1 ⫽ mild, 2 ⫽ moderate, and 3 ⫽ severe); pupil and corneal
the visit at which the investigator decided to switch/discontinue the
diameters were measured; and ophthalmoscopy was performed.
subject—a decision made after the IOP was measured—was used
Visual acuity, biomicroscopy, and IOP measurement were re-
to impute the week 12 IOP using the last observation carried
peated using criteria and procedures established at screening.
forward method.
Pachymetry was performed in subjects able to cooperate.
For the primary end point, an analysis of covariance (ANCOVA)
Before randomization, eligible subjects were stratified by age
model with treatment and baseline diagnosis as factors and
(0 –⬍3, 3–⬍12, and 12–18 years), diagnosis (PCG or non-PCG, baseline IOP as a covariate was used to analyze the difference in
including secondary glaucomas and JOAG), and baseline IOP IOP reduction at week 12 between latanoprost and timolol. The
value of the study eye (⬍27, 27–31, and ⬎31 mmHg). If both eyes treatment difference (latanoprost – timolol) and its 95% confidence
qualified, the eye with the highest baseline IOP measurement was interval (CI) were calculated, and latanoprost was considered
designated as the study eye; if baseline IOP measurements were noninferior, was either more or similarly effective, to timolol if the
the same in 2 qualified eyes, the right eye was designated as the lower limit of the 95% CI of the difference was ⬎⫺3 mmHg
study eye. Within each stratum, subjects were randomized 1:1 to (noninferiority margin agreed to by the European Medicines
latanoprost versus timolol using a centralized system administered Agency Pediatric Committee as a component of the Pediatric
by a third party that maintained control of subject assignment Investigator Plan). The analysis plan provided for an option of
codes. Subjects assigned to the latanoprost group instilled 1 drop switching to superiority in the event that the lower limit of the 95%
of latanoprost 0.005% once daily at 8 PM (⫾30 minutes) and 1 drop CI for the treatment difference not only lay above the noninferi-
of the latanoprost vehicle (placebo) once daily at 8 AM (⫾30 ority margin (⫺3 mmHg) but also above zero.
minutes). Subjects assigned to receive timolol 0.5% (or optionally The sample size of 120 subjects (60 subjects in each treatment
0.25% for subjects aged ⬍3 years) twice daily instilled 1 drop at arm) was estimated to provide an 84% power to demonstrate that
8 AM (⫾30 minutes) and 1 drop at 8 PM (⫾30 minutes). The first latanoprost was not inferior to timolol within a margin of 3 mmHg,
dose of study medication was to be administered at the baseline assuming latanoprost had 1 mmHg reduction more than timolol in
visit in the morning after randomization, and the last dose was mean IOP change from baseline, a common standard deviation
instilled at 8 AM of the week 12/early discontinuation visit. (SD) of 7 mmHg, and a 10% dropout rate. The same sample size
Follow-up visits were scheduled at weeks 1, 4, and 12. If at would provide an 84% power for a superiority test to detect a
weeks 1 or 4, the investigator determined that the IOP was not difference of 4 mmHg between the latanoprost and timolol treat-
controlled in a subject depending on each subject’s condition and ment arms with the same assumption of a common SD of 7 mmHg
treatment history, the subject could be switched to open-label and a 10% dropout rate.
concomitant therapy of latanoprost 0.005% at 8 PM (⫾30 minutes) As recommended by the International Conference on Harmoni-
and timolol 0.5% (or optionally 0.25% for subjects aged ⬍3 years) sation E9 guideline22 and the European Agency for the Evaluation
twice daily at 8 AM (⫾30 minutes) and 8 PM (⫾30 minutes). If such of Medicinal Products Points to Consider guidance document,23
a change was made, the subject returned for additional follow-up the primary noninferiority efficacy analysis was based on the per
at weeks 2 or 5; if the IOP remained uncontrolled, the subject protocol (PP) population; a supporting analysis was based on the
discontinued study participation. Intraocular pressure measure- intent-to-treat (ITT) population. The primary efficacy analysis
ments at each follow-up visit occurred at 10 AM (⫾1.5 hours). population for the determination of superiority would be based on
Preferably, the same assessor, masked to subject assignment dur- the ITT population. The PP population was restricted to subjects

3
 Ophthalmology Volume xx, Number x, Month 2011

Figure 1. Subject disposition.

with no major protocol violations who received at least 1 week of baseline conjunctival hyperemia score and the highest score at
study medication and who had an IOP measurement at week 1. The any post-baseline visit was calculated to reflect the maximum
ITT population was defined as all subjects randomized into the shift in grade.
study who received at least 1 dose of study medication.
Secondary efficacy end points included the mean change from
baseline in the study eye at each visit (weeks 1, 4, and 12), the Results
mean IOP level at each visit (baseline and weeks 1, 4, and 12), and
the proportion of subjects with at least a 15% IOP reduction from A total of 139 subjects were randomized to study treatment at 42
baseline at both weeks 4 and 12 (responders). For analyses of centers (80 subjects at 24 EU centers and 59 subjects at 18 non-EU
secondary end points, missing data were not imputed, and only centers; Fig 1). Although 69 subjects were assigned to receive
IOP levels measured while a subject was on masked monotherapy latanoprost and 70 subjects were assigned to receive timolol, 1
were analyzed. An ANCOVA model with treatment and baseline subject in each randomization group did not start study medication
diagnosis as factors and baseline IOP as a covariate was used to because they were no longer willing to participate in the study.
estimate the difference in IOP reduction at each visit between Subjects discontinuing during the study included 4 of 68 (5.9%) in
latanoprost and timolol; corresponding 95% CIs also were calcu- the latanoprost-treated group versus 8 of 69 (11.6%) in the timolol-
lated. The statistical significance of between-treatment differences treated group (not significant [NS]). Half (6/12) of the discontinu-
in proportion of responders was evaluated using the Cochran- ations occurred in the youngest subjects (0 –⬍3 years); discontinu-
Mantel-Haenszel chi-square test stratified by baseline diagnosis ations were noted for 3 subjects each in the age groups 3 to ⬍12
(PCG vs. non-PCG). years and 12 to 18 years. All 137 treated subjects were included in
Efficacy analyses also were performed for the PCG and non- the ITT and safety populations, whereas 107 of 137 (78.1%)
PCG subgroups. For these analyses, the ANCOVA model included subjects were included in the PP population.
only treatment as the factor and baseline IOP as a covariate. For Baseline demographic and diagnostic characteristics for the
the responder analysis, a Pearson’s chi-square test was used to ITT population are summarized in Table 2 (available at http://
evaluate the treatment difference. aaojournal.org). More than three quarters of subjects in both treat-
Adverse events were classified by body system and preferred ment groups were white, and approximately half were male. The
term using the Medical Dictionary for Regulatory Activities distribution across age groups was similar between treatments. The
(MedDRA Version 12.1) coding system. Safety end points were primary diagnosis for approximately 45% of subjects in each
analyzed using the ITT population. The difference between the group was PCG. The most common non-PCG diagnosis was

4
 Maeda-Chubachi et al 䡠 Latanoprost vs Timolol in Pediatric Glaucomas

JOAG. Both eyes were treated in 33 of 62 (53.2%) subjects with

PCG, in 38 of 40 subjects (95.0%) with JOAG, and in 2 of 17
(11.8%) subjects with either aphakic or pseudophakic glaucoma.
The proportion of subjects in the latanoprost group who had
received ocular pharmacologic treatments for glaucoma before the
start of the study was 69% compared with 55% of subjects in the
timolol group (difference NS); these prior treatments most com-
monly included latanoprost, timolol, and brinzolamide. The latano-
prost- and timolol-treated subjects were similar with respect to
previous glaucoma-related surgery (reported for 17/68 [25.0%] vs.
16/69 [23.2%] subjects, respectively) and any other prior ocular
surgery (reported for 23/68 [33.8%] vs. 22/69 [31.9%] subjects,
respectively).
In the ITT population, the median duration of treatment, in-
cluding open-label therapy, for both groups was 85.0 days (range:
latanoprost, 7–92 days; timolol, 2–105 days). The median duration
of masked monotherapy was 84.0 days in the latanoprost group Figure 3. Mean ⫾ standard error IOP reduction from baseline (mmHg) by
(range: 6 –92 days) and 85.0 days in the timolol group (range: study visit, study eye, PP population excluding observations collected after
2–105 days). Similar durations of treatment were noted in the PP switching to open-label concomitant therapy. IOP ⫽ intraocular pressure;
population. PP ⫽ per protocol.
Both latanoprost and timolol reduced mean IOP levels in the
study eye from baseline to week 12 with least square mean
reductions of 7.2 and 5.7 mmHg, respectively (Table 3, available groups. Mean IOP reductions at week 12 were numerically greater
at http://aaojournal.org). The least square mean difference was 1.5 in the latanoprost group than in the timolol group, and this differ-
mmHg in favor of latanoprost, and the 95% CI of the least square ence was more evident in the non-PCG subgroup compared with
mean difference was between ⫺0.81 mmHg (0.81 mmHg in favor the PCG subgroup. The lower bounds of the 95% CI of the
of timolol) and 3.74 mmHg (3.74 mmHg in favor of latanoprost; difference between the least square means in both diagnosis sub-
P⫽0.21). The lower limit of the 95% CI of the difference (⫺0.81 groups were ⬎⫺3 mmHg (Fig 2). Response rates in the PCG
mmHg) was ⬎⫺3 mmHg, which demonstrated that latanoprost subgroup were 50% (14/28) and 46% (12/26) for latanoprost- and
was not inferior (i.e., was either more or similarly effective) to timolol-treated subjects, respectively (difference NS); in the non-
timolol (Fig 2). Mean IOP reductions in the latanoprost group were PCG groups, response rates were 72% (18/25) and 57% (16/28),
numerically but not statistically significantly greater than in the respectively (difference NS).
timolol group at weeks 1, 4, and 12 (Fig 3). In all, 60% of subjects Both latanoprost and timolol were well tolerated. Greater pro-
(32/53) treated with latanoprost and 52% of subjects (28/54) portions of timolol- than latanoprost-treated subjects reported
treated with timolol were responders (ⱖ15% IOP reduction from treatment-related adverse events and serious adverse events and
baseline at weeks 4 and 12; P⫽0.33; PP population). discontinued treatment because of adverse events (Table 5). Seri-
Results of the analysis of the primary end point by diagnosis ous adverse events considered by the investigator to be possibly
(PCG vs. non-PCG) are summarized in Table 4 (available at related to study drug treatment included uncontrolled developmen-
http://aaojournal.org). In general, non-PCG subjects had higher tal glaucoma and eye hemorrhage in 1 timolol-treated subject and
mean baseline IOP levels and numerically greater IOP reductions pneumonia in 1 timolol-treated subject. Treatment-related con-
from baseline to week 12 than the PCG subjects in both treatment junctival hyperemia judged to be an adverse event by an investi-
gator in the study eyes was reported as none in latanoprost-treated
subjects and 3 (4.3%) in timolol-treated subjects. Baseline hyper-
emia was noted in 10 and 17 study eyes in the latanoprost and
timolol groups, respectively; worsening from baseline, defined as
a 1 grade change on a hyperemia grading score at each visit,
occurred in 12 of 68 (17.6%) subjects in the latanoprost group and
in 10 of 69 (14.5%) subjects in the timolol group. In study eyes, 1
treatment-related case each of punctate keratitis, abnormal eye
sensation, and photophobia was reported in the timolol group, and
1 such case of installation site pain was reported in those treated
with latanoprost. No clinically relevant changes were noted in
other ocular safety measures (i.e., visual acuity, corneal thickness,
biomicroscopy, ophthalmoscopy, pachymetry, or gonioscopy) or
in systemic safety measures (i.e., vital signs, alertness, or physical
examination results).

Discussion

Figure 2. Latanoprost noninferior to timolol. In this study, lower limit of

Because uncontrolled glaucoma in young children may lead
the CI of the difference was considered to indicate (A) inferiority if ⱕ⫺3, to a lifetime of blindness,24,25 it is important to find safe and
(B) noninferiority if ⬎⫺3 and ⱕ0, or (C) superiority if ⬎0. CI ⫽ effective treatments for this condition. Our study demon-
confidence interval; IOP ⫽ intraocular pressure; ITT ⫽ intent-to-treat; strated that in pediatric patients with glaucoma, the adult
PCG ⫽ primary congenital glaucoma; PP ⫽ per protocol. dose of latanoprost (0.005% ophthalmic solution once

5
 Ophthalmology Volume xx, Number x, Month 2011

Table 5. All Causality Treatment-Emergent Adverse Events, Intent-to-Treat Population

Latanoprost (Nⴝ68) Timolol (Nⴝ69)

No. of adverse events 41 64
Subjects with adverse events, n (%) 24 (35.3) 29 (42.0)
No. of treatment-related adverse events 6 15
Subjects with treatment-related adverse events, n (%) 4 (5.9) 10 (14.5)
Subjects with serious adverse events, n (%) 2 (2.9) 7 (10.1)
Subjects discontinued because of adverse events, n (%) 1 (1.5) 4 (5.8)
Subjects discontinued because of treatment-related adverse 0 1 (1.5)
events, n (%)
Died during the study 0 0

n ⫽ number of subjects included in the assessment; N ⫽ number of subjects in the population.

daily) was not inferior (i.e., was either more or similarly to the trabecular outflow pathway and, thus, slightly less
effective) to timolol 0.5% (or optionally 0.25% for those responsive to latanoprost.16
aged ⬍3 years) administered twice daily in its ability to Although the effect of outflow modulating pharmacother-
reduce IOP. There was a trend favoring greater IOP reduc- apy in PCGs may not be straightforward, there may be value in
tion with latanoprost; however, latanoprost superiority was latanoprost being used as a “last resort” for cases that are
not found on statistical analysis. There was no statistically inadequately controlled with other drugs or before undergoing
significant difference in responder rates (proportion of sub- surgical treatment.14 In the PCG subpopulation of the present
jects with a ⱖ15% IOP reduction) between latanoprost and study, the IOP-lowering effect was 5.9⫾5.9 mmHg with la-
timolol, with approximately half of all patients having a tanoprost compared with 5.3⫾4.2 mmHg with timolol.
treatment response. Although 4 patients receiving latanoprost discontinued
Results of previous retrospective studies suggested that the study, twice as many patient discontinuations occurred
patients with PCG have little clinically relevant IOP- in the timolol treatment group. There were no treatment-
lowering response with pharmacologic agents such as la- related discontinuations due to latanoprost or its lack of
tanoprost, brinzolamide, and levobetaxolol.14,26 In the pres- efficacy compared with 4% of patients receiving timolol.
ent study, the IOP-lowering effect of latanoprost was greater Likewise, latanoprost was well tolerated with a low inci-
in the non-PCG subjects than in the PCG subjects. Approx- dence of adverse events over the 3 months of exposure. In
imately one half of the subset of non-PCG subjects had a both treatment groups, the youngest subjects (0 –⬍3 years)
diagnosis of JOAG, with the remaining diagnoses including experienced a relatively higher frequency of adverse events
unspecified secondary glaucomas, aphakic glaucoma, and compared with the 2 older age groups (3–⬍12 years and
Sturge–Weber syndrome. In contrast with previously re- 12–18 years), suggesting that the youngest children are
ported results, in this study, latanoprost did demonstrate a more likely to exhibit overt nonocular infection-related
clinically relevant IOP-lowering effect in the PCG subgroup symptoms such as rhinitis, pneumonia, and gastroenteritis.
(mean 5.9 mmHg; P⬍0.01), although the observed IOP The rates of conjunctival hyperemia were low in both
reductions suggest the magnitude of IOP-lowering effect groups, but the proportion of patients or parents reporting a
may be less than in the non-PCG subgroup (mean 8.4 worsening of the conjunctival hyperemia score was greater
mmHg; P⬍0.01) in general. The differences between these in the latanoprost group. It seems most likely that cases of
findings in the PCG group in this present study and previous conjunctival hyperemia were not considered by the inves-
smaller and retrospective studies may be explained by re- tigators as being of clinical significance and were not re-
gional practice differences. In particular, in the retrospective ported as adverse events.
studies conducted in the United States, latanoprost often Major limitations of this study were the recruitment
was added as an adjunctive therapy to patients with PCG challenges and the conduct in this pediatric population. In
initially being managed with surgery. This may have se- particular, the rarity of glaucoma in this age group made it
lected out refractory cases of glaucoma in which the uncon- challenging to enroll patients. Consequently, statistical dif-
trolled IOP required additional medical therapy; thus, the ferences in IOP reduction between latanoprost and timolol
true efficacy of latanoprost was blunted because of the were difficult to determine given the small numbers of
underlying nature of the disease. patients in the study, especially in the very young children.
The reasons for a smaller IOP-lowering response in This most likely also contributed to the lack of ability to
patients with PCG overall are not entirely clear. Latanoprost further explore the relative effects of latanoprost versus
acts by increasing the uveoscleral outflow via relaxation of timolol as monotherapy for pediatric glaucomas of different
the ciliary muscle27 and remodeling of connective tissue in causes or subtypes. Another limitation was the nonstandard-
the trabecular meshwork between the outer longitudinal ized method of identification of ocular structural changes in
muscle bundles.28 In eyes with PCG (or other forms of terms of glaucomatous progression.
glaucoma that predominate in children), the uveoscleral In conclusion, latanoprost seemed to be safe and effec-
outflow pathway may be anatomically abnormal in addition tive for the subjects in this trial over a 3-month exposure

6
 Maeda-Chubachi et al 䡠 Latanoprost vs Timolol in Pediatric Glaucomas

period, supporting the use of this pharmacologic agent as a 14. Black AC, Jones S, Yanovitch TL, et al. Latanoprost in
treatment option for patients with pediatric glaucoma. Fur- pediatric glaucoma—pediatric exposure over a decade. J AA-
thermore, approximately half of the subjects with PCG had POS 2009;13:558 – 62.
a treatment response, suggesting that latanoprost also may 15. Enyedi LB, Freedman SF, Buckley EG. The effectiveness of
latanoprost for the treatment of pediatric glaucoma. J AAPOS
have a role in the treatment of PCG. 1999;3:33–9.
Acknowledgments. The authors thank Charles Bosworth, PhD,
16. Enyedi LB, Freedman SF. Latanoprost for the treatment
for contributions to the conduct of the study, including its concep-
of pediatric glaucoma. Surv Ophthalmol 2002;47(Suppl):
tion and design. Editorial support, including contributing to the
S129 –32.
first draft of the article, revising after author comments, and styling
17. Yang CB, Freedman SF, Myers JS, et al. Use of latanoprost in
the article for journal submission, was provided by Jane G. Mur-
the treatment of glaucoma associated with Sturge-Weber syn-
phy, PhD, of Zola Associates and was funded by Pfizer Inc.
drome. Am J Ophthalmol 1998;126:600 –2.
18. Altuna JC, Greenfield DS, Wand M, et al. Latanoprost in
glaucoma associated with Sturge-Weber syndrome: benefits
References and side-effects. J Glaucoma 1999;8:199 –203.
19. Ong T, Chia A, Nischal KK. Latanoprost in port wine stain
1. Mandal AK, Netland PA. Terminology and Classification of related paediatric glaucoma. Br J Ophthalmol 2005;87:
Developmental Glaucoma. In: Mandal AK, Netland PA, eds. 1091–3.
The Pediatric Glaucomas. Philadelphia, PA: Butterworth 20. Regulation (EC) No. 1901/2006 of the European Parliament
Heinemann; 2006:5–9. and of the Council, of 12 December 2006, on medicinal
2. Papadopoulos M, Khaw PT. Childhood glaucoma. In: Taylor products for paediatric use and amending Regulation (EEC)
D, Hoyt CS, eds. Pediatric Ophthalmology and Strabismus. No. 1768/92, Directive 2001/20/EC, Directive 2001/83/EC,
3rd ed. New York: Elsevier Saunders; 2005:458 –71. and Regulation (EC) No. 726/2004. Available at: http://eur-
3. Papadopoulos M, Cable N, Rahi J, Khaw PT, BIG Eye lex.europa.eu/LexUriServ/site/en/consleg/2006/R/02006R1901-
Study Investigators. British Infantile and Childhood Glau- 20070126-en.pdf. Accessed August 25, 2010.
coma (BIG) Eye Study. Invest Ophthalmol Vis Sci 2007; 21. Pediatric Research Equity Act of 2003. Pub L No. 108-155,
48:4100 – 6. 117 Stat 1936-1943. Available at: http://www.glin.gov/view.
4. Taylor RH, Ainsworth JR, Evans AR, Levin AV. The epide- action?glinID⫽175685#. Accessed August 25, 2010.
miology of pediatric glaucoma: the Toronto experience. J 22. International Conference on Harmonisation; guidance on sta-
AAPOS 1999;3:308 –15. tistical principles for clinical trials; availability—FDA. No-
5. Terraciano AJ, Sidoti PA. Management of refractory glau- tice. Fed Regist 1998;63(179):49583–98. Available at: http://
coma in childhood. Curr Opin Ophthalmol 2002;13:97–102. frwebgate.access.gpo.gov/cgi-bin/getdoc.cgi?dbname⫽1998_
6. Marquis RE, Whitson JT. Management of glaucoma: focus on register&docid⫽98-24754-filed.pdf. Accessed February 25,
pharmacological therapy. Drugs Aging 2005;22:1–21. 2011.
7. Realini T, Fechtner RD. 56,000 ways to treat glaucoma. Oph- 23. European Agency for the Evaluation of Medicinal Products
thalmology 2002;109:1955– 6. Committee for Proprietary Medicinal Products (CPMP).
8. Moore W, Nischal KK. Pharmacologic management of glau- Points to consider on switching between superiority and non-
coma in childhood. Paediatr Drugs 2007;9:71–9. inferiority. London: EMEA; 2000. Available at: http://www.
9. Papadopoulos M, Khaw PT. Advances in the management of tga.gov.au/docs/pdf/euguide/ewp/048299en.pdf. Accessed Au-
paediatric glaucoma. Eye (Lond) 2007;21:1319 –25. gust 25, 2010.
10. Distelhorst JS, Hughes GM. Open-angle glaucoma. Am Fam 24. Levy J, Tessler Z, Tamir O, Lifshitz T. Primary congenital
Physician 2003;67:1937– 44. glaucoma [in Hebrew]. Harefuah 2004;143:876 – 80, 910.
11. van der Valk R, Webers CA, Schouten JS, et al. Intraocular 25. Biglan AW. Glaucoma in children: are we making progress? J
pressure-lowering effects of all commonly used glaucoma AAPOS 2006;10:7–21.
drugs: a meta-analysis of randomized clinical trials. Ophthal- 26. Whitson JT, Roarty JD, Vijaya L, et al, Brinzolamide Pediatric
mology 2005;112:1177– 85. Study Group. Efficacy of brinzolamide and levobetaxolol in
12. Alm A, Schoenfelder J, McDermott J. A 5-year, multicenter, pediatric glaucomas: a randomized clinical trial. J AAPOS
open-label, safety study of adjunctive latanoprost therapy for 2008;12:239 – 46.
glaucoma. Arch Ophthalmol 2004;122:957– 65. 27. Alm A. Prostaglandin derivatives as ocular hypotensive
13. Goldberg I, Li XY, Selaru P, Paggiarino D. A 5-year, ran- agents. Prog Retin Eye Res 1998;17:291–312.
domized, open-label safety study of latanoprost and usual care 28. Toris CB, Gabelt BT, Kaufman PL. Update on the mechanism
in patients with open-angle glaucoma or ocular hypertension. of action of topical prostaglandins for intraocular pressure
Eur J Ophthalmol 2008;18:408 –16. reduction. Surv Ophthalmol 2008;53(Suppl):S107–20.

Footnotes and Financial Disclosures

4
Originally received: September 10, 2010. NIHR Biomedical Research Centre for Ophthalmology, Moorfields Eye
Final revision: March 1, 2011. Hospital and UCL Institute of Ophthalmology, London, United Kingdom.
Accepted: March 3, 2011. 5
University of Utah, Moran Eye Center, Salt Lake City, Utah.
Available online: ●●●. Manuscript no. 2010-1247. 6
Pfizer Inc, New York, New York.
1
Pfizer Inc, New London, Connecticut. Presented at: the World Ophthalmology Congress, June 5–9, 2010, Berlin,
2 Germany; and in part at: the Annual Meeting of the 2010 American
Private Practice, Jupiter, Florida.
Academy of Ophthalmology, October 16 –19, 2010, Chicago, Illinois.
3
Duke Eye Center, Durham, North Carolina. Financial Disclosure(s):

7
 Ophthalmology Volume xx, Number x, Month 2011

The author(s) have made the following disclosure(s): TM-C, KC-B, and analysis, and interpretation of the data; and in the preparation, review, and
EY are employees of Pfizer Inc. BW was an employee of Pfizer Inc at the approval of the manuscript. Peng T. Khaw, MD, was funded by the NIHR
time the study was conducted and the article was developed. BW, BDS, Biomedical Research Centre for Ophthalmology at Moorfields Eye Hos-
and SFF are currently consultants for Pfizer Inc. PTK is currently a pital and UCL Institute of Ophthalmology.
consultant for Pfizer Inc, Alcon, Allergan, Bausch & Lomb, and AstraZen-
eca. Correspondence:
Funding: This study was sponsored by Pfizer Inc, New York, New York. Tomoko Maeda-Chubachi, MD, PhD, MBA, Pfizer Inc, 50 Pequot Avenue,
The sponsor participated in the design of the study; in the management, B3240, New London, CT 06320. E-mail: [email protected].