The Relationship Between Statin Use andOpen-Angle Glaucoma

Joshua D. Stein, MD, MS,1 Paula Anne Newman-Casey, MD,1 Nidhi Talwar, MA,1 Bin Nan, PhD,2

Julia E. Richards, PhD,1,3 David C. Musch, PhD, MPH1,3

Purpose: To determine whether 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitors (statins) affectthe risk of developing open-angle glaucoma (OAG) in persons with hyperlipidemia.

Design: Retrospective, longitudinal cohort analysis.Participants: Individuals aged �60 years with hyperlipidemia enrolled in a national United States managed

care network between 2001 and 2009.Methods: Multivariable Cox regression analyses were performed to assess the relationship between statin

use and the development of OAG (from no prior OAG diagnosis), progression from a prior diagnosis of glaucomasuspect to a diagnosis of OAG, and need for medical or operative interventions for OAG. Regression models wereadjusted for sociodemographic factors and medical and ocular comorbidities.

Main Outcome Measures: Hazard ratios (HRs) with 95% confidence intervals (CIs).Results: Of the 524 109 individuals with hyperlipidemia, 316 182 (60%) had �1 outpatient prescription for

statins. The hazard of developing OAG decreased 0.3% (adjusted HR, 0.997; 95% CI 0.994–0.999) for everyadditional month of statin consumption. Individuals with hyperlipidemia who took statins continuously for 2 yearshad an 8% (adjusted HR, 0.922; 95% CI, 0.870–0.976) decreased OAG risk relative to those who received nostatin therapy. The hazard of progressing from a diagnosis of glaucoma suspect to OAG decreased 0.4%(adjusted HR, 0.996; 95% CI, 0.993–0.999) for every additional month of statin exposure. Individuals who tookstatins continuously for 2 years had a 9% (adjusted HR, 0.907; 95% CI, 0.846–0.973) decreased risk ofprogressing from glaucoma suspect to OAG relative to those who received no statin therapy. The hazard ofrequiring medical treatment for OAG decreased 0.4% (adjusted HR, 0.996; 95% CI, 0.993–0.998) for everyadditional month of statin exposure. No differences in need for glaucoma surgery were noted among those withOAG who were and were not taking statins (adjusted HR, 1.002; 95% CI, 0.994–1.010).

Conclusions: Statin use was associated with a significant reduction in the risk of OAG among persons withhyperlipidemia. Given the mounting evidence of statin protection against OAG including both basic science andobservational clinical studies, an interventional prospective study might provide additional insights into the roleof statins in the prevention of early OAG.

Financial Disclosure(s): The authors have no proprietary or commercial interest in any of the materialsdiscussed in this article. Ophthalmology 2012;119:2074–2081 © 2012 by the American Academy of Ophthalmology.

ph

wSOIpaowdcOs

Statins (3-hydroxy-3-methylglutaryl-coenzyme A reductaseinhibitors) are a class of medications widely used to lowercholesterol in patients with hyperlipidemia. Recent evidencesuggests that statins are capable of reducing the risk of cere-brovascular and cardiovascular events, independent of theireffect on cholesterol levels, which has resulted in an expansionof the indications for their usage.1,2 There has been growingevidence in the basic science and clinical literature supportingthe notion that statins may also be useful in patients withseveral diseases of the central nervous system, including isch-emic stroke, Alzheimer’s disease, and multiple sclerosis.3 Be-cause the optic nerve and retinal nerve fiber layer are theprimary structures damaged in patients with open-angle glau-coma (OAG), evidence that statins play a protective role inpatients with other diseases of the central nervous system leads

us to hypothesize that these agents may also be beneficial in o

2074 © 2012 by the American Academy of OphthalmologyPublished by Elsevier Inc.

reventing the development of OAG in persons who are atigh risk for this visually disabling condition.

There are conflicting findings in the literature as tohether statins may be beneficial in patients with OAG.ome studies suggest that statins may be protective againstAG,4–6 whereas other studies do not support this effect.7,8

n a recent study assessing the relationship between com-onents of metabolic syndrome and OAG, we found that,lthough adults with diabetes mellitus or hypertension dem-nstrated an increased risk of developing OAG, enrolleesith hyperlipidemia actually had a 5% decreased risk ofeveloping OAG.9 Our earlier work also demonstrated thatomorbid hyperlipidemia attenuated the increased risk ofAG associated with diabetes or hypertension. In this

tudy, we seek to determine whether it is the hyperlipidemia

r the medications used to treat hyperlipidemia (statins) that

ISSN 0161-6420/12/$–see front matterhttp://dx.doi.org/10.1016/j.ophtha.2012.04.029

Naloc

DTmfmst3c3wa

AS9rci

hcoyldbtwlhppfpcal

tdtidai2todcemsI

w

Stein et al � Statins and Open-angle Glaucoma

may be responsible for the reduced risk of OAG that wasobserved in the earlier study.

Using a large, nationwide healthcare claims database con-taining detailed medical records for �500 000 older Ameri-cans with hyperlipidemia, we sought to determine whether anassociation exists between statin use and the development ofOAG, the conversion from glaucoma suspect to OAG, or theneed for medical or surgical therapy for OAG. Should statinsbe found to be protective against OAG; this may lead to novelprevention strategies for this visually disabling condition.

Methods

Data SourceThe i3 InVision Data Mart database (Ingenix, Eden Prairie, MN)contains detailed records of all beneficiaries in a managed carenetwork with members throughout the United States. The datasetcontains all individuals with �1 International Classification ofDiseases, Ninth Revision (ICD-9-CM)10 codes for eye-related di-agnoses (360–379.9), �1 Current Procedural Terminology11

codes for any eye-related visits, diagnostic, or therapeutic proce-dures (65091–68899 or 92002–92499), or any other claim submit-ted by an ophthalmologist or optometrist from January 1, 2001,through December 31, 2009. For each enrollee, we had access toall medical claims for ocular and nonocular conditions and so-ciodemographic information including age, sex, race, educationlevel, and household net worth. The database also contained re-cords of all outpatient pharmacy prescriptions that were filled.During the time period when a beneficiary was enrolled in themedical plan, they were also fully enrolled in the pharmacy plan.This database has been used in the past to study patients withOAG.9,12,13

Because all the data were deidentified, the University of Mich-igan’s Institutional Review Board determined that this study wasexempt from requiring their approval.

Participants and Sample SelectionIndividuals were included in the analysis if they met the followingcriteria: age �60 years, continuous enrollment in the medical planfor �2 years, �2 visits to an eye care provider (ophthalmologist oroptometrist), and �1 diagnosis of hyperlipidemia (ICD-9-CMcodes 272.0–272.9).

Statin UseIndividuals were classified as receiving statins if they had �1outpatient pharmacy prescription for any of the following medi-cations: lovastatin (Altocor, Altoprev, Mevacor), cerivastatin(Baycol), amlodipine besylate/atorvastatin calcium (Cauduet), ro-suvastatin (Crestor), fluvastatin (Lescol, Lescol XL), atorvastatin(Lipitor), pravastatin (Pravachol), simvastatin (Zocor), PravigardPAC (pravastatin/buffered aspirin), niacin plus lovastatin (Advi-cor), or ezetimibe and simvastatin (Vytorin). The database con-tains information on the number of days for which an enrollee wassupplied a given medication and this information was included inthe regression model (described later). In addition to identifyingand quantifying statin use, we also assessed exposure to the fol-lowing nonstatin cholesterol-lowering medications: fenofibrate(Antara, Fenoglide, Fibricor, Lipofen, Lofibra, Tricor, Triglide), clo-fibrate (Atromid-s), cholestyramine (Locholest, Locholest light, Pre-valite, Questran, Questran light), cholestyramine light, cholestyramine

resin, colestid (Colestipol HCL), gemfibrozil (Lopid), niacin (Niacor, s

iaspan, Slo-niacin), fenobric acid (Trilipix), colesevelam (Welchol),nd ezetimibe (Zetia). Individuals who were prescribed niacin plusovastatin or ezetimibe and simvastatin were counted in the statin andther cholesterol-lowering medication groups because these productsontain both statin and nonstatin cholesterol-lowering medications.

ependent Variableshe dependent variables for these analyses included the develop-ent of OAG (from no prior OAG diagnosis), the progression

rom a diagnosis of glaucoma suspect to one of OAG, the need foredical therapy for OAG, and the need for laser or incisional

urgery for OAG. Beneficiaries were identified as having OAG ifhey had �1 of the following ICD-9-CM codes: 365.1, 365.10,65.11, 365.12, and 365.15 during their time in the plan. Thoseoded with glaucoma suspect only (ICD-9-CM codes 365.0,65.00, 365.01, 365.04) who never received a diagnosis of OAGere considered to be glaucoma suspect only, not OAG (Table 1,

vailable at http://aaojournal.org).

nalysestatistical analyses were performed using SAS software, version.2 (SAS Inc, Cary, NC). Participant characteristics were summa-ized for the entire sample using means and standard deviations forontinuous variables and frequencies and percentages for categor-cal variables.

Cox regression with delayed entry was used to estimate theazard of developing OAG (going from no record of ICD-9-CModes 365.1, 365.10, 365.11, 365.12, and 365.15 to a diagnosis ofne of these codes) associated with statin use. We used the first 2ears each beneficiary was enrolled in the medical plan as aook-back period. Individuals who received �1 OAG diagnosisuring this look-back period were excluded from the analysisecause we were unable to determine with any certainty whetherhey had the condition before enrollment in the medical plan orhether they were first diagnosed with this condition during the

ook-back period. To be even more certain that enrollees did notave the outcome of interest (i.e., OAG), during this look-backeriod we also required �1 visit to an eye care provider during thiseriod with no record of this condition. All beneficiaries wereollowed in the model from the index date (2 years after entry into thelan) until they either developed the outcome of interest or wereensored. Censoring occurred at the time of a participant’s last visit ton eye care provider during the follow-up period because this wasikely their last opportunity to receive the event of interest.

Our key predictor variable was the use of statins, which wasreated as a time-dependent covariate in the models. The number ofays each beneficiary had covered by a prescription for statins wasotaled over a 24-month period for every day they were followedn the plan before the event or censoring. This amount of use, inays, was then converted to months for ease of interpretation. Inddition, we also estimated the change in hazard for each outcomef statins were taken for a year in the last 2 years or for the entire-year period. Multivariable models were adjusted for age (theime axis), sex, race, education level, household net worth, regionf residence at the time of medical plan enrollment, time-depen-ent use of nonstatin cholesterol-lowering medications, ocularomorbidities (cataract, pseudophakia or aphakia, macular degen-ration, diabetic retinopathy), medical comorbidities (diabetesellitus, systemic hypertension, obesity, systemic hypotension,

leep apnea, migraine headache), and the Charlson Comorbidityndex (a measure of overall health).14

Additional Cox regression models were performed to assesshether statin use was associated with progression from glaucoma

uspect to OAG (going from an ICD-9-CM diagnosis of glaucoma

2075

sowr(3

iicri(lm8tsv

idenc

Ophthalmology Volume 119, Number 10, October 2012

suspect [365.00, 365.01, or 365.04] to the receipt of an ICD-9-CMcode of OAG [365.1, 365.10, 365.11, 365.12, and 365.15]), theneed for a topical glaucoma medication (prostaglandin analogues,�-adrenergic receptor blockers, �-2 adrenergic agonists, carbonicanhydrase inhibitors, and parasympathomimetics), and, amongindividuals who were known to have OAG, the need for laser orincisional glaucoma surgery.

Finally, a sensitivity analysis was performed to determinewhether the association between the main predictor variables andOAG was affected if we changed our definition of OAG fromrequiring �1 ICD-9-CM code for OAG to requiring �2 codes forOAG. We also tested an interaction between age and statin use onrisk of developing OAG. For all analyses, P�0.05 was consideredsignificant.

Results

There were 524 109 persons with hyperlipidemia who met the study

Total Beneficiaries Who Received Any Form of Eye Care from

2001-2009

N= 10,326,832

Beneficiaries Eligible for Analysis

N = 682,525

Beneficiaries with 1 Diagnosis of Hyperlipidemia

N = 524,109

Glaucoma Suspects Converting to OAG Analysis

Excluded cases* (N=474,481)

Need for GlauMedication A

Excluded cases* (N

Beneficiaries Eligible for Analysis

N=49,628

Beneficiaries EliAnalysis

N=241,25

Figure 1. Selection of beneficiaries for the analysis. IOP � intraocular prthe outcome, no visit to an eye care provider either in the lookback periono evidence of glaucoma suspect and for glaucoma surgery analysis no ev

inclusion criteria. (Fig 1). The cohort of beneficiaries who met inclu- 3

2076

ion criteria were enrolled in the plan for a mean � standard deviationf 4.3�1.8 years. Among those enrolled in the study, the mean ageas 68.1�6.7 years. There were 280 543 females (53.5%) and the

acial distribution included 413 066 whites (87.4%), 28 004 blacks5.9%), 19 651 Latinos (4.2%), 8644 Asian Americans (1.8%), and497 individuals of other races (0.7%; Table 2).

A total of 316 182 out of the 524 109 enrollees with hyperlip-demia (60.3%) had �1 prescription for a statin during their timen the plan, 20 776 (4%) enrollees were prescribed only nonstatinholesterol-lowering medications, and 187 151 (35.7%) beneficia-ies were not prescribed any cholesterol-lowering medication dur-ng the follow-up period. Among the 316 182 statin users, 92 95529.4%) also filled �1 prescription for a nonstatin cholesterol-owering medication. For those who were receiving statins, theean length of time they were taking these medications was

00�621 days (range, 1–3266 days) and 96% used these medica-ions for �30 days’ duration. The most commonly prescribedtatins were atorvastatin and simvastatin. The median number ofisits to eye care providers for users and nonusers of statins was

Excluded Beneficiaries

1. Noncontinuous enrollment in the medical and/or pharmacy plan

2. Enrollment in the plan for 2 years

3. Age 60 years

4. Beneficiaries without one of more CPT code indicating a visit to an ophthalmologist or optometrist

New Diagnosis of OAG Analysis

Excluded cases* (N=282,398)

s

,858)

Need for Glaucoma Surgery Analysis

Excluded cases* (N=515,873)

for Beneficiaries Eligible for Analysis N=8236

Beneficiaries Eligible for

Analysis N=241,711

Beneficiaries Excluded Because they Did Not Have Hyperlipidemia

N = 158,416

; OAG � open-angle glaucoma. *Excluded due to a nonincident case ofn the follow-up period for glaucoma suspect converting to OAG analysise of incident OAG to begin with.

comanalysi

=282

gible 1

essured or i

.0.

abt

CAt(dag9s20r2dO

Stein et al � Statins and Open-angle Glaucoma

Development of OAG

Overall, 10 266 individuals (4.3%) received �1 incident OAGdiagnosis during their time in the medical plan, out of the 241 711who met the eligibility criteria. The mean age of those whodeveloped incident OAG was 69.7�6.7 years compared with68.9�6.7 years for those who did not develop OAG. Blacks,Latinos, and Asian Americans had an increased hazard of devel-oping OAG (P�0.05) relative to whites.

After adjustment for confounding factors, the hazard of devel-oping OAG decreased 0.3% (adjusted hazard ratio [HR], 0.997;95% confidence interval [CI], 0.994–0.999; P � 0.0056) for everyadditional month of statin consumption. Persons with hyperlipid-emia who took statins for 1 year over the prior 2 years had a 4%decreased hazard (adjusted HR, 0.960; 95% CI, 0.933–0.988) ofOAG relative to individuals with hyperlipidemia who never re-ceived statins in the last 2 years, assuming all other characteristicswere the same. Likewise, persons who took statins continuouslyfor 2 years had an 8% decreased hazard (adjusted HR, 0.922; 95%CI, 0.870–0.976) of OAG relative to those who did not receive

Table 2. Demographic Charac

Development of OAGAnalysis

ConversionGlaucoma Su

OAG An

BeneficiariesWho Did NotDevelop OAG

BeneficiariesWho

DevelopedOAG

BeneficiariesWho Did

Not Convertfrom

GlaucomaSuspect to

OAG

Beneficiaries, n 231 445 10 266 42 694Age (yrs)

Mean � SD 68.9�6.7 69.7�6.7 68.0�6.6Range 60–86.8 60–86 60–87

Gender, n (%)Male 100 576 (43.5) 4514 (44.0) 18 115 (42.4)Female 130 869 (56.5) 5752 (56.0) 24 579 (57.6)

Race, n (%)White 187 456 (89.3) 7875 (84.5) 33 457 (86.5)Black 10 240 (4.9) 740 (8.0) 2414 (6.2)Latino 7305 (3.5) 437 (4.7) 1657 (4.3)Asian-American 3353 (1.6) 201 (2.2) 836 (2.2)Other race 1491 (0.7) 61 (0.6) 309 (0.8)

OAG � open-angle glaucoma; SD � standard deviation.

Event*BeneficiariIncluded,

New diagnosis of OAG 241 711Conversion from glaucoma suspect to OAG 49 628New prescription of IOP-lowering medication 241 251Need for glaucoma laser or incisional surgery 8236

CI � confidence interval; HR � hazard ratio; IOP � intraocular pressur*Those diagnosed with the event during the 2-year look-back period wer†Reference group: Patients not on statins at all in the past year.‡Reference group: Patients not on statins at all in the past 2 years.

§For example, a hazard ratio of 0.996 means a 0.4% decreased risk of developin

ny statins in the last 2 years (Table 3). There was no associationetween consumption of nonstatin cholesterol-lowering medica-ions and development of OAG (P � 0.12).

onversion from Suspected Glaucoma to OAGfter excluding individuals who were diagnosed with OAG during

heir first 2 years in the health plan, there were 6934 beneficiaries14.0%) who developed OAG among the 49 628 who had beeniagnosed in the look-back period with glaucoma suspect. Afterdjustment for confounding factors, the hazard of progressing fromlaucoma suspect to OAG decreased 0.4% (adjusted HR, 0.996;5% CI, 0.993–0.999; P � 0.0062) for every additional month oftatin exposure. Persons who took statins for 1 year over the pastyears had a 5% decreased hazard (adjusted HR, 0.952; 95% CI,

.920–0.986) of progressing from glaucoma suspect to OAGelative to glaucoma suspects who never received statins in the lastyears. Those who took statins continuously for 2 years had a 9%

ecreased hazard (adjusted HR, 0.907; 95% CI, 0.846–0.973) ofAG relative to the group who did not receive any statins in the

tics of Beneficiary Population

to Need for GlaucomaMedications Analysis

Need for Glaucoma SurgeryAnalysis

ciarieshoertedeing acomaect to

G

BeneficiariesWho Were Not

NewlyPrescribed aGlaucomaMedication

BeneficiariesWho Were

NewlyPrescribed aGlaucomaMedication

BeneficiariesWho Did Not

NeedGlaucoma

Surgery

BeneficiariesWho Needed

GlaucomaSurgery

34 229 831 11 420 7227 1009

�6.6 69.0�6.8 69.7�6.7 69.8�6.7 69.5�6.386 60–86.8 60–86 60–86 60–84.5

(43.5) 99 535 (43.3) 4969 (43.5) 3184 (44.1) 431 (42.7)(56.5) 130 296 (56.7) 6451 (56.5) 4043 (55.9) 578 (57.3)

(84.8) 185 306 (88.9) 8873 (85.0) 5566 (85.0) 799 (85.3)(7.6) 10 986 (5.3) 761 (7.3) 526 (8.0) 66 (7.0)(4.9) 7265 (3.5) 489 (4.7) 282 (4.3) 45 (4.8)(2.1) 3378 (1.6) 239 (2.3) 136 (2.1) 18 (1.9)(0.6) 1460 (0.7) 79 (0.7) 41 (0.6) 9 (1.0)

Table 3. Protective

Beneficiaries withIncident Event, n

HR (95% CI) per Month ofStatin Use; P Value

10 266 0.997 (0.994–0.999)§; 0.00566934 0.996 (0.993–0.999); 0.0062

11 420 0.996 (0.993–0.998); 0.00021009 1.002 (0.994–1.010); 0.6811

G � open-angle glaucoma.luded to exclude nonincident cases.

teris

fromspect

alysis

BenefiW

Convfrom bGlauSusp

OA

69

69.360–

30183916

533548130613337

esn

e; OAe exc

g OAG with every additional month of statin use.

2077

S

Sra((tmysi

D

Iiiwdrotttorgyw9tlmmrctnptOo

Ophthalmology Volume 119, Number 10, October 2012

last 2 years (Table 3). There was no difference in the hazard ofprogressing from glaucoma suspect to OAG for those who usednonstatin cholesterol-lowering medications (P � 0.077).

Need for Medical Therapy for OAG

There were 47511 enrollees (17.0%) who were prescribed �1glaucoma medication during their time in the plan. After excludingany beneficiary who was treated with glaucoma medications dur-ing their first 2 years in the plan, there were 11420 enrollees whowere newly prescribed a glaucoma medication. After adjustmentfor confounders, the hazard for being prescribed a glaucoma med-ication decreased 0.4% (adjusted HR, 0.996; 95% CI, 0.993–0.998; P � 0.0002) for every additional month of statin exposure.Plan participants who took statins for 1 year over the past 2 yearshad a 5% decreased hazard (adjusted HR, 0.950; 95% CI, 0.924–0.976) of requiring an intraocular pressure–lowering medicationrelative to those who never received statins in the last 2 years.Taking statins continuously for 2 years resulted in a 10% de-creased hazard (adjusted HR, 0.902; 95% CI, 0.854–0.953) ofrequiring a pressure-lowering medication relative to never takingany statins in the last 2 years (Table 3). The hazard for beingprescribed a glaucoma medication decreased 0.6% (adjusted HR,0.994; 95% CI, 0.990–0.998; P � 0.0017) per month for thoseindividuals treated with nonstatin cholesterol-lowering medica-tions. Persons who took nonstatin cholesterol-lowering medica-tions for 1 year over the past 2 years had a 7% decreased risk(adjusted HR, 0.928; 95% CI, 0.886–0.972) of being prescribedglaucoma medications and if a person took a nonstatin cholesterol-lowering medication for 2 years, they had a 14% (adjusted HR,0.862; 95% CI, 0.785–0.946) decreased risk of being prescribed aglaucoma medication.

Need for Surgical Interventions for OAG

Among the 8236 persons with incident OAG who had no glau-coma surgical intervention coded before their incident diagnosis ofOAG, 1009 (12.3%) went on to require laser or incisional glau-coma surgery during their time in the plan. The average length oftime from first diagnosis of OAG to laser or incisional glaucomasurgery was 268�387 days. After adjustment for confounders, thehazard of an individual with OAG later requiring laser or inci-sional glaucoma surgery was not significantly different (adjustedHR, 1.002; 95% CI, 0.994–1.010) with each additional month ofstatin exposure (P � 0.68; Table 3). We found a similar result fornonstatin cholesterol-lowering medications (adjusted HR, 0.992;95% CI, 0.978–1.006; P � 0.27).

Effects of Statins

Decreased Hazard ofDeveloping Event per Month

of Medication Use (%)HR (95% CI) per Year of

Statin Use; P Value

DecreasWi

Contin

0.3 0.960 (0.933–0.988)§; 0.00560.4 0.952 (0.920–0.986); 0.00620.4 0.950 (0.924–0.976); 0.0002

1.021 (0.925–1.127); 0.6811

2078

ensitivity Analyses

ensitivity analyses were performed to determine whether theesults changed if we required a confirmatory diagnosis of OAG,nd the findings did not materially change from those presentedresults not shown). In addition, we performed descriptive analysesTable 4, available at http://aaojournal.org) and tested an interac-ion between age and statin use in the multivariable regressionodel to determine whether statin use by older individuals (�67

ears old) affects the risk of developing OAG differently fromtatin use among younger individuals (�67 years old) and thenteraction was not significant; P � 0.43.

iscussion

n this longitudinal analysis of a large group of older Amer-cans with hyperlipidemia, after adjustment for a number ofmportant confounding factors, we found that those whoere prescribed statins experienced a decreased risk ofeveloping OAG from no prior OAG diagnosis, a decreasedisk of converting from a diagnosis of glaucoma suspect tone of OAG, and a decreased risk of receiving �1 prescrip-ion for intraocular pressure—lowering medications. Fur-hermore, a dose–response effect was observed, wherebyhe greater the number of months of exposure to statinsver the past 2-year duration, the greater the reduction inisk of developing OAG or requiring medical treatment forlaucoma. Individuals who were prescribed statins for 1ear had a 5% decreased hazard of developing OAG,hereas those who were prescribed statins for 2 years had a% decreased hazard of OAG relative to those who were notaking statins. For those taking nonstatin cholesterol-owering medications, we found a reduction in the use ofedical therapy for glaucoma, but we did not find that theseedications conferred a significant reduction in risk of

eceiving a diagnosis of OAG, of converting from a glau-oma suspect to OAG or requiring glaucoma surgery. Al-hough one of the mechanisms through which statins andonstatin cholesterol-lowering medications may exert theirrotective effect is through lowering cholesterol, the facthat statins significantly attenuated the hazard of developingAG and of converting from glaucoma suspect to OAG andther cholesterol-lowering medications did not suggests that

sk of EventYear ofStatin Use†

HR (95% CI) per 2 Yearsof Statin Use; P Value

Decreased Risk of EventWith 2 Years of

Continuous Statin Use‡

0.922 (0.870–0.976)§; 0.0056 8%0.907 (0.846–0.973); 0.0062 9%0.902 (0.854–0.953); 0.0002 10%1.042 (0.856–1.269); 0.6811

ed Rith 1uous

4%5%5%

bstfdcatpanwt

S

TtrssinraomLMasAMacbb

eccnwOdoseifiafibUcsFh

Stein et al � Statins and Open-angle Glaucoma

statins may have an effect on the development of glaucomathat is independent of their cholesterol-lowering properties.

Several other population-based studies and studies usingadministrative claims data have investigated the relationshipbetween statin use and glaucoma. In a case-control study ofmale patients at a Veterans Administration Medical Center(VAMC), McGwin et al4 found that individuals prescribedstatins for �24 months had 40% reduced odds of develop-ing OAG. This group also showed a protective effect againstOAG among those prescribed other cholesterol-loweringagents.4 A prospective study by Leung et al5 demonstratedthat statin use was associated with visual field stabilizationover 3 years among patients with normal tension glaucoma.5

De Castro et al6 showed that statin use slowed glaucoma-tous changes to the optic nerve and nerve fiber layer onconfocal scanning laser ophthalmoscopy. Although thesestudies, like ours, suggest that statins may be protectiveagainst OAG, not all do. A large case-control study byOwen et al,7 using information from a primary care databasein the United Kingdom found no relationship between statinuse and OAG. Using administrative data from Canada,Iskedjian et al8 showed no difference in the need for addi-tional pressure-lowering medications among persons takingprostaglandin analogs with or without statins. Direct com-parisons of these studies with one another and with thefindings of the present analysis are challenging becauseof differences in study design, differences in the extent towhich potential confounding factors were accounted for,and differences in the sociodemographic profiles of thestudy patients. Because hyperlipidemia is associated withother components of the metabolic syndrome, includinghypertension and diabetes mellitus— both of which havebeen found to be associated with OAG15–27—whenstudying the relationship between statin use and OAG, itis important to adjust for these medical comorbidities.Some of the prior studies did not consider these potentialconfounders, unlike the current study, which did.

There are several proposed mechanisms that could ex-plain why individuals who are taking statins have a reducedrisk of OAG. Statins have been shown to upregulate endo-thelial nitric oxide synthase. Nitric oxide synthase causesvasodilation and an increase in retinal and choroidal bloodflow.28 Improved retinal and choroidal perfusion may beimportant in maintenance of the health of the optic nerveand retinal nerve fiber layer. In addition, statins are knownto exert a neuroprotective effect in the setting of ischemiaand may protect retinal ganglion cells in a similar way, aswas demonstrated in an ischemia–reperfusion model of therat retina.29 Proposed mechanisms for the neuroprotectiveeffect of statins include decreasing gluatamate-mediatedcytotoxicity30,31 and protecting against apoptosis in the cen-tral nervous system.3 There is some evidence that increasedplasma nitric oxide lowers intraocular pressure.32,33 Statinsmay lower intraocular pressure via other mechanisms aswell; research has shown that statins affect various molec-ular intermediaries in the aqueous outflow pathways, in-cluding rho kinase activity and myosin II adinosine triphos-phatase activity, which all increase aqueous outflow facilitythrough the trabecular meshwork and result in a reduction of

intraocular pressure, a known risk factor for glaucoma.34,35 n

A secondary aim of this analysis was to identify, to theest of our abilities and within the limitations of a retro-pective review of claims data, whether statins are protec-ive against OAG early or later in the disease course. Weound that statins conferred a protective effect against theevelopment of OAG from no prior OAG diagnosis, theonversion from a diagnosis of glaucoma suspect to OAG,nd the need for a medical therapy for OAG, suggesting thathese medications may be beneficial early in the diseaserocess. To evaluate whether statins are beneficial in moredvanced glaucoma, we also assessed differences in theeed for glaucoma surgery among persons with OAG whoere and were not taking statins and found no reduction in

he need for surgery among statin users (P � 0.68).

trengths and Limitations

he strengths of this study include its large sample size andhe ability for us to follow patients over time to assess theelationship between cholesterol-lowering medication expo-ure and subsequent development of OAG. Unlike othertudies of this subject that were conducted at specific med-cal centers, the patients in this study come from commu-ities throughout the United States. In addition, instead ofelying on patient self-report to identify medical conditionsnd exposure to medications, the data from this study werebtained from providers and pharmacy records, and thusay be more accurate (Patty LE, Wu S, Torres M, Varma R.ALES Group, Doheny Eye Institute/Keck- USC School ofedicine. Validity of self-reported diagnosis and treatment

mong Latinos in the Los Angeles Latino Eye Study. Posteression 516. Correlates and Outcomes of Eye Diseases.ssociation of Research and Vision in Ophthalmology,ay 6, 2010, Fort Lauderdale, Florida.) Finally, in this

nalysis we were able to adjust for a variety of potentialonfounding factors, including other components of meta-olic syndrome, which have previously been documented toe associated with OAG.

There are several limitations that need to be acknowl-dged. The data source used in this analysis does not containlinical information, such as intraocular pressure, centralorneal thickness, or findings from visual field and opticerve evaluations. Therefore, we cannot tell with certaintyhether all of the beneficiaries who were diagnosed withAG indeed had this condition, nor could we fully captureisease severity to incorporate into our analyses. Second,ur analysis did not consider results of laboratory testing,uch as levels of different types of cholesterol and triglyc-rides. Third, because all of the enrollees in this plan hadnsurance, one must be cautious in generalizing our studyndings to other populations, such as uninsured individualsnd those residing outside the United States. That our studyndings are consistent with findings from the research doney McGwin et al,4 which was conducted at a VAMC in thenited States, and the study by Leung et al,5 which was

onducted in China, suggest that the relationship betweentatin use and OAG may be applicable to other groups.ourth, it is possible that enrollees who take statins are moreealth conscious than others with hyperlipidemia who are

ot prescribed these medications, and this may impact their

2079

1

1

1

1

1

1

1

1

1

1

2

2

2

Ophthalmology Volume 119, Number 10, October 2012

use of eye care services. If these individuals have greateruse of eye care services, we would expect this to increase,not decrease, their risk of receiving an OAG diagnosis. Onthe other hand, those who have untreated hyperlipidemiamay have more severe ocular comorbidities, such as dia-betic retinopathy, which would necessitate more eye careservices and increase their chance of being diagnosed withOAG. In this analysis, we sought to deal with this issue bycontrolling for other ocular comorbidities, although theremay be additional confounders not adequately captured inclaims data. Fifth, we have no way of knowing with cer-tainty whether enrollees actually consumed all of the pre-scribed medications. The presumption in this type of anal-ysis is that a patient will not keep filling prescriptions if theyare not taking the drug. Thus, although these data cannotprovide assurance that a specific patient took a prescribedstatin drug on a specific date, the expectation is that thosewho repeatedly filled prescriptions were actually taking theprescribed statin medication as opposed to those who werenever prescribed the statin medication, and that a tally ofthe number of days for which prescriptions were filled is areasonable surrogate measure of drug consumption. In ad-dition, poor adherence with medications, if anything, shouldbias our findings to the null, and yet we are finding statinsto be protective against OAG.

Finally, it is important to acknowledge that our findingsmay not be generalizable to individuals who do not havehyperlipidemia. Statins may not exert a protective effectagainst developing OAG if the person does not have aderangement in their lipid metabolism. However, in thestudy of glaucoma patients at a VAMC by McGwin et al,4

statins had a protective effect against developing OAG inpatients with hyperlipidemia, as well as for the subset ofpersons who were taking statins for other reasons besidestreatment of hyperlipidemia.

The findings of this analysis support earlier studies dem-onstrating that treating hyperlipidemia with a statin plays aprotective role in patients with OAG, especially in the earlystages of the disease. After adjustment for important con-founding influences, we found that enrollees taking statinshad a decreased hazard of developing OAG, of progressingfrom glaucoma suspect to OAG, and of requiring pressure-lowering medications. Moreover, our study shows a dose–response effect of statin exposure whereby the longer anenrollee was prescribed these medications, the greater theprotective effect. Given the mounting evidence of statinprotection against OAG, including both basic science andobservational clinical studies, an interventional prospectivestudy might provide additional insights into the role ofstatins in the prevention of early OAG.

References

1. Sillesen H, Amarenco P, Hennerici MG, et al, SPARCL Inves-tigators. Atorvastatin reduces the risk of cardiovascular events inpatients with carotid atherosclerosis: a secondary analysis of theStroke Prevention by Aggressive Reduction in Cholesterol Levels(SPARCL) trial. Stroke 2008;39:3297–302.

2. Downs JR, Clearfield M, Weiss S, et al, AFCAPS/TexCAPS

Research Group. Primary prevention of acute coronary events

2080

with lovastatin in men and women with average cholesterollevels: results of AFCAPS//TexCAPS. JAMA 1998;279:1615–22.

3. Schmeer C, Kretz A, Isenmann S. Statin-mediated protectiveeffects in the central nervous system: general mechanisms andputative role of stress proteins. Restor Neurol Neurosci 2006;24:79–95.

4. McGwin G Jr, McNeal S, Owsley C, et al. Statins and othercholesterol-lowering medications and the presence of glau-coma. Arch Ophthalmol 2004;122:822–6.

5. Leung DY, Li FC, Kwong YY, et al. Simvastatin and diseasestabilization in normal tension glaucoma: a cohort study.Ophthalmology 2010;117:471–6.

6. De Castro DK, Punjabi OS, Bostrom AG, et al. Effect of statindrugs and aspirin on progression in open-angle glaucomasuspects using confocal scanning laser ophthalmoscopy. ClinExperiment Ophthalmol 2007;35:506–13.

7. Owen CG, Carey IM, Shah S, et al. Hypotensive medication,statins, and the risk of glaucoma. Invest Ophthalmol Vis Sci2010;51:3524–30.

8. Iskedjian M, Walker JH, Desjardins O, et al. Effect of selectedantihypertensives, antidiabetics, statins, and diuretics on ad-junctive medical treatment of glaucoma: a population basedstudy. Curr Med Res Opin 2009;25:1879–88.

9. Newman-Casey PA, Talwar N, Nan B, et al. The relationshipbetween components of metabolic syndrome and open-angleglaucoma. Ophthalmology 2011;118:1318–26.

0. International Classification of Diseases. 9th revision, ClinicalModification. Physician ICD-9-CM 2006. Vol 1. Chicago:AMA Press; 2006.

1. CPT 2006: Current Procedural Terminology. Professional ed.Chicago: AMA Press; 2006.

2. Stein JD, Kim DS, Niziol LM, et al. Differences in rates ofglaucoma among Asian Americans and other racial groups,and among various Asian ethnic groups. Ophthalmology2011;118:1031–7.

3. Stein JD, Kim DS, Mundy KM, et al. The association betweenglaucomatous and other causes of optic neuropathy and sleepapnea. Am J Ophthalmol 2011;152:989–98.

4. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A newmethod of classifying prognostic comorbidity in longitudi-nal studies: development and validation. J Chronic Dis1987;40:373– 83.

5. Kahn HA, Milton RC. Alternative definitions of open-angleglaucoma: effect on prevalence and associations in the Fra-mingham Eye Study. Arch Ophthalmol 1980;98:2172–7.

6. Klein BE, Klein R, Jensen SC. Open-angle glaucoma andolder-onset diabetes: the Beaver Dam Eye Study. Ophthalmol-ogy 1994;101:1173–7.

7. Mitchell P, Smith W, Chey T, Healey PR. Open-angle glau-coma and diabetes: the Blue Mountain Eye Study, Australia.Ophthalmology 1997;104:712–8.

8. Chopra V, Varma R, Francis BA, et al, Los Angeles LatinoEye Study Group. Type 2 diabetes mellitus and the risk ofopen-angle glaucoma: the Los Angeles Latino Eye Study.Ophthalmology 2008;115:227–32.

9. Katz J, Sommer A. Risk factors for primary open angleglaucoma. Am J Prev Med 1988;4:110–4.

0. Pasquale LR, Kang JH, Manson JE, et al. Prospective study oftype 2 diabetes mellitus and risk of primary open-angle glau-coma in women. Ophthalmology 2006;113:1081–6.

1. Bonovas S, Peponis V, Filioussi K. Diabetes mellitus as a riskfactor for primary open-angle glaucoma: a meta-analysis. Dia-bet Med 2004;21:609–14.

2. Zafra Perez JJ, Vellegas Perez MP, Canteras Jordana M,

Miralles De Imperial J. Intraocular pressure and prevalence of

3

3

3

3

3

3

Stein et al � Statins and Open-angle Glaucoma

occult glaucoma in a village of Murcia [in Spanish]. Arch SocEsp Oftalmol 2000;75:171–8.

23. Bulpitt CJ, Hodes C, Everitt MG. Intraocular pressure andsystemic blood pressure in the elderly. Br J Ophthalmol 1975;59:717–20.

24. Dielemans I, Vingerling JR, Algra D, et al. Primary open-angle glaucoma, intraocular pressure, and systemic bloodpressure in the general elderly population: the RotterdamStudy. Ophthalmology 1995;102:54–60.

25. Tielsch JM, Katz J, Sommer A, et al. Hypertension, perfusionpressure, and primary open-angle glaucoma: a population-based assessment. Arch Ophthalmol 1995;113:216–21.

26. Bonomi L, Marchini G, Marraffa M, et al. Vascular riskfactors for primary open angle glaucoma: the Egna-NeumarktStudy. Ophthalmology 2000;107:1287–93.

27. Wilson MR, Hertzmark E, Walker AM, et al. A case-controlstudy of risk factors in open angle glaucoma. Arch Ophthal-mol 1987;105:1066–71.

28. Nagaoka T, Takahashi A, Sato E, et al. Effect of systemicadministration of simvastatin on retinal circulation. Arch Oph-thalmol 2006;124:665–70.

29. Honjo M, Tanihara H, Nishijima K, et al. Statin inhibits

leukocyte-endothelial interaction and prevents neuronal death

Footnotes and Financial Disclosures

March 4, 2011.

FTm

SAcdt

CJ1

induced by ischemia-reperfusion injury in the rat retina. ArchOphthalmol 2002;120:1707–13.

0. Zacco A, Togo J, Spence K, et al. 3-Hydroxy-3-methylglutarylcoenzyme A reductase inhibitors protect cortical neurons fromexcitotoxicity. J Neurosci 2003;23:11104–11.

1. Bosel J, Gandor F, Harms C, et al. Neuroprotective ef-fects of atorvastatin against glutamate-induced excitotox-icity in primary cortical neurones. J Neurochem 2005;92:1386 –98.

2. Kotikoski H, Oksala O, Vapaatalo H, Aine E. Aqueous humourflow after a single oral dose of isosorbide-5-mononitrate inhealthy volunteers. Acta Ophthalmol Scand 2003;81:355–60.

3. Kotikoski H, Vapaatalo H, Oksala O. Nitric oxide and cyclicGMP enhance aqueous humor outflow facility in rabbits. CurrEye Res 2003;26:119–23.

4. Zhang M, Rao PV. Blebbistatin, a novel inhibitor of myosin IIATPase activity, increases aqueous humor outflow facility inperfused enucleated porcine eyes. Invest Ophthalmol Vis Sci2005;46:4130–8.

5. Song J, Deng PF, Stinnett SS, et al. Effects of cholesterol-lowering statins on the aqueous humor outflow pathway. In-

vest Ophthalmol Vis Sci 2005;46:2424–32.

Originally received: November 22, 2011.Final revision: April 18, 2012.Accepted: April 27, 2012.Available online: June 21, 2012. Manuscript no. 2011-1680.1 Department of Ophthalmology and Visual Sciences, University of Mich-igan Medical School, Ann Arbor, Michigan.2 Department of Biostatistics, University of Michigan School of PublicHealth, Ann Arbor, Michigan.3 Department of Epidemiology, University of Michigan School of PublicHealth, Ann Arbor, Michigan.

Presented in part at: the American Glaucoma Society Annual Meeting

inancial Disclosure(s):he authors have no proprietary or commercial interest in any of theaterials discussed in this article.

upported by the National Eye Institute K23 Mentored Clinician Scientistward (J.D.S.: 1K23EY019511-01); American Glaucoma Society Clini-

ian Scientist Grant (J.D.S.), Blue Cross Blue Shield of Michigan Foun-ation (J.D.S.), Research to Prevent Blindness (D.C.M. and J.E.R.); Na-ional Eye Institute Core Grant EY00703.

orrespondence:oshua D. Stein, MD, MS, University of Michigan, Kellogg Eye Center,

000 Wall Street, Ann Arbor, MI 48105. E-mail: jdstein@med.umich.edu.

2081