Chapter 2 An Evidence-Based Approach to Glaucoma Care

Louis R. Pasquale

Put several glaucoma specialists in a room and solicit opinions on just about any clinical issue and you are bound to receive multiple opinions on how to mailage the problem. For example, a patient presents \tith shallow angles on van Herrick screening. What gonioscopic features would prompt you to perform a laser iridotomy? Would you perforni a dark room provocative test to confirni your impression that the angle was potentially occludable? Do you perforni a pharma- cological challenge test before proceeding to prophylactic laser iridotoniy? JVould you order an ultrasound biomicroscopic test to confirm your impression? If you do decide that prophylactic laser iridotomy is indicated, what is the best surgical technique for achieving patency? Is there evidence to support a "best approach" to the patient with the narrow angle? In this chap- ter, we discuss how evidence-based medicine (EBM) can be used to provide the very best answers to questions such as these. This approach can be applied not only to glaucoma problems, but also to any clinical problem in medicine.

Yet there is considerable confusion regarding EBM. An arti- cle in Enze magazine stated that EBM was "a hard, cold empirical look at what wolks, what doesn't and how to dis- tinguish between the This statement about EBhI suk- gab that all patient problems have straightforward solution& and EBM helps find those solutions fairly fead$ly. Clearly, wa r?omodpemoulntw glaueoma problem's that are not straighdowand at all.~+For instance, should we commit a European-derived Caucasian male patient with thin corneas, intraocular pressure (TOP) in the high-teens, glaucoma-like discs, and unreliable visual field findings to medical therapy for glaucoma? The Oouba~ [Hyperten'sSon Treatment Stu* ,(,QHT&l, a randomized conml trial"(RC!'T) compating TOP 1 0 w ~ ~ ~ ~ @ r a ~ y versus obsmation.in~thhtupreVthtion of pri- "

~ r m q open-angle glauooma (POAO) in patients with ocul& .hypartens.ion (OHTN), found thinner central corneal thick- ness (CCT) to be a risk factor for conversion to POAG among patients with OHTN4; but while our patient has thin CCT, his IOP is normal. The Barbados Eye Study5 found that thinner CCT was an independent risk factor for POAG in a popula-

2.1 An Introduction to Evidence-Based tion of African descent, where subjects had s spectrum of IOP ranging from normal to high; llowever, this patient is

Medicine Caucasian, which raises the question of whether the data are

The era of evidence-based medicine is upon us. The Cochrane Collaboration (a large collection of volunteers who use a sys- tematic approach to evaluate the efficacy and safety of medi- cal interventions)' has outmuscled the senior, distinguished, and experienced collection of dinner speakers (the so-called members of "en~inence-based medicine")? whisked in from dar by pharmaceutical coinpanies to tell us what is best for our patients. Today, even when "thought leaders'ya buzz- word that finds its origin i n the business world used to denote someone held in high regard because of his or her trendset- ting ideas or the appearance of his or her names on mail-in sunreys) are recruited to speak about a medical product by industry, continuing medical education guidelines dictate that they must provide a balanced and evidence-based presentation.

specifically applicable to his situation. Furthermore, there are absolutely no data to suggest that such treatment would be cost-effective at this time. Finally, there are patient-spe- cific issues that may make the answer to this question even more difficult to arrive at. For example, if the patient is 80 years old and has a pill-rolling tremor secondary to Parkinson's disease and is already taking eight other oral agents for other systemic illnesses, then the physician may opt to observe the patient. On the other hand, if the patient is 53 years old, expresses anxiety about glaucoma blindness, and has a strong family history for the disease, then the physician may be inclined to initiate treatment.

Dr. David Sackett, often regarded as the father of EBM, defines EBM as "the conscientious, explicit and judicious use of current best evidence in making decisions about the

I P.N. Schacknow and J.R Samples (eds.), The Glaucoma Book: A Pracrical, El~idence-BasedApproach to Patient Care, 23 DO1 10.1P07l978-0-387-76700-0-2, O Springer Science+Business Media, U C 2010

24 L.R. Pasauale -- - -

care of individual patient^."^ EBM is a process and rhere is patient comfortable, and theoretically it should enhance nothing cold and calculating about it. The process starts with resurfacing of the cornea epithelial defect - it had to be the a patient in your exam lane. npical ly that patient has a correct approach! That night, the glaucoma specialist did a clinical problem that cannot be answered by consulting an litemturesearch on the management of cornea abrasions, and ophthalmology textbook or a single paper in the literature, to his amazement he found a meta-analysis published in unless that paper happens to be an EBM review on the topic. 1998 synthesizing the results of five randomized clinical tri- For example, a 45-year-old African-American female with a als comparing the patch versus no patch approach to manag- positive family history of glaucoma blindness presents wit11 ing traumatic corneal abrasions. The meta-analysis concluded intraocular pressure (IOP) of 24 rnnlHg OU, central corneal that there was no advantage to patching as long as the patient thicknesses of 545 pm OU, and CDR=0.4, OU. A standard was treated with a topical antibiotic, topical cycloplegia, and automated perimetry test is performed and is reliable and topical nonsteroidill anti-inflammatory agent.' In fact, two normal in both eyes. Do we treat this patient? If we do treat randomized clinical trials (RCTs) published after this sys- her, what would be the best initial approach? Fixst, we 5; tematic literature review reached similar conclusions.8 Of assemble what is called h e "evidancs cart$" Admittedly, course, it did not mean patching corneal abrasions was akin the cart may not be loaded with items (as it might be if we to malpractice; but it did indicate that this glaucoma special- were dealing with breast cancer or myocardial infarction), ist who completed his ophthnlrnology residency in 1990 and but the items that are there may be quite appropriate for was not treating corneal abrasions on a regular basis any- that patient sitting in your exam lane. Them, we need to- Wore did not stay up-to-date in the mmagement of traumatic w a p e , t h e quality of .the euidanoe i n that ~wbpEhaJly, ws corneal abrasion. He was not aware that one could manage a~p1y @at evidence to our patient a&best as we cm, given cornenl abrasions without resorting to patching the eye prior the medical and nonmedical circumstances related to her to delivering a lecture of the subject in year 2000. Oh by the case. Later in this chapter, we will demonstrate the EBbI way, I was that glauconla specialist and that was the "ah-hay process in detail. moment that sparked my interest in EBM.

2.2 Evidence-Based Medicine: An Interesting 2.2.7 The Evolution o f Evidence-Based Story Medicine

In the year 2000, a glaucoma specialist was asked to give a talk on basic eye emergencies directed to primary care physi- cians. For a glaucoma specialist with a tertiary level glau- coma referral practice, delivering a talk on this subject seemed like a relatively easy task. During the talk, the glau- coma specialist pointed out that the best way to manage a traumatic coi~leal abrasion was to instill an antibiotic oint- ment and Cyclogyl 1% drops topically, pressure patch the eye, and follow-up with the patient the next day. A very clear rationale for this approach was provided: The antibiotic pro- vided coverage against microbial invasion of the corneal stroma during the time when there was a breach in the over- lying epithelial barrier; the cycloplegic agent reduced uveal spasm induced by the highly innervated cornea; and the pres- sure patch immobilized the lid to allow cornenl limbal stem cells to resurface the epithelial break and settle onto the underlying Bowman's membrane. How could that possibly be the wrong way to manage garden-variety corneal abra- sions? At the end of the lecture, one of the course attendees politely approached the glaucoma specialist and stated that he might not be practicing EBM and that most abrasions could be treated without patching. This sounded like heresy to the glaucoma specialist. After all, the patch made the

While the origin of EBM can be traced back to medieval times,I0 the trends in medical care that resulted in the current embrace of EBM can be traced back to the American Revolution, starting withaphysician-signer of the Declaration of Independence, Dr. Benjamin Rush. Dr. Rush's approach to disease was to zealously attack and conquer it. After all, it was this philosophy that allowed the colonists to ultimately prevail over a seemingly formidable British adversary. Unfortunately, there were no real tools to address most of the medical problems encountered; and the attacks fervently employed by Dr. Rush (limited exsanguinations, the induc- tion of emesis, blistering, etc.) typically took a "one-size- fits-all" approach, without evidence that they actually helped patients." Ultimately Dr. Rush's ardent medical philosophy was supplanted by the realization that most medical thera- pies of the time were largely ineffective, The highly influen- tial Sir William Osler popularized this latter philosophy. His approach to disease can be summed up in one of his more famous aphorisms: One ofthe first duties of the physician is to educate the mosses not to take rnedici~zes,~' Sir Osler, regarded as the father of modem medicine, stated that the role of the physician was to provide accurate diagnoses, prognoses, and palliative care for most conditions. Beginning

2 An Evidence-Based Approach to Glaucoma Care 25

DO everylhing- Glaucoma Prevention Study (EGPS).I6 The EGPS, was a placebo-controlled RCT that compared dorzolarnide (a topical carbonic anhydrase inhibitor) versus placebo in the prevention of POAG among patients with OIJTN. The EGPS has similar characteristics to OHTS, allowing investigators to pool the data from both trials in estimating the risk of developing POAG. The latest calculator represents a significant revision

Do nothing - fiom the earlier tools in that it does not use diabetes status in

I estimating the risk of converting from OHTN to POAG. The

I I ' - 7 0 0 1700's 1900's 1940's initial versions of the calculator assigned a reduced risk of

Dr. Rush Dr. Qsler D ~ . I;.scteR conversion from OHTN to POAG among diabetic patients because the OHTS found that a self-report of diabetes mel-

Fig. 2.1 The medical profession's approach to treatment ~ntenlentions litus was associated with a reduced risk of from over time

OHTN to POAG. However, when data from OHTS were combined with the EGPS, diabetes was no longer related

in the early 1940s, the field of medicine witnessed the antibiotic revolution, the introduction of chemotherapy agents, and the discovery of prednisone. These agents looked like miraculous substances compared to what treatments were available to address medical conditions in the early 1900s (see IvIorrris" for ahistorical account of the impact of antibiotic therapy in the 1940s). Nonetheless, initial enthusi- asm for these new tools led to their indiscriminate and often ill-advised use that did not necessarily translate into clinical benefit for our patients. f~&d'-~&4@908qtQc~~Qa~d~Pe!&bD b ~ d u e e d - % h ~ f idea thfh.ztthWE ghurfla w~lbw~&e.~fmmre d%'mLEyk. ' 1 6 ~ ~ ~ V f n f l ~ n ~ f i t wt!$r evftlenosto @d@t~fdf&l

.,s;lseiaiom&@The relation between the degrees of inter- vention for medical illness as a function of time dating from the late 1700s to present is illustrated in Fig. 2.1.

with conversion from OHTN to POAG. Hence, the latest cal- culator provides the most updated assessment of risk. Certainly, more refinements of the glaucoma risk calculator are fortl~coming, such as adjustment for life expectancy, and if you practice EBM, you will be familiar with these refine- ments because EBM engages you in following new develop- ments in managing disease.

qaw,t~,g?~hppP&@U is that .learning the prin- cia M wiP mi&@ yon a perpe;tu&l smdent of medi-

and, after all, was not lifelong learning about disease one of the reasons we entered this noble profession in the first place? You can begin practicing EBM now, You can review the medical decision-making that you make during the course of a day in your office and follow the EBM pro- cess to see what the consensus of evidence suggests i best way to arrive at a particular management decision ~ b ~ ~ ~ ~ ~ & ~ l a ' ~ & ~ @ ~ r ~ ' & w ~ ~ ~ ~ b h a ~ l i t e m ~ ~ ~ d o e ~ ' '

~~~L~~~yous:us:~~~~~epmbl~~~~. While one may view this as a shortcoming of the EBM approach;hmayiTI4b

2.2.2 Evidence-Based Medicine: Why Bother? rwmwppp &P ,GFkl OmGsgar +<LO Y,BnOa &- i d ~ i h ~ w L ~ e ~ ' h ~ ~ p l ~ a u l a m e a . -r

There are compelling reasons to use the EBM process in the There are other benefits to practicing EBM. h@YW 7on.a -'-

delivery of health care. First, there is an increasing volun~e of sm&~g&=md~~hww&~urd~~de- what ~question~ me evidence available to address clinical problems, and the EBM ~ p a ~ ~ d ~ ~ ~ ~ & e c a u s e they involve your patients) approach helps keep the clinician abreast of this evidence in and you can find out the answer to these questions (if there a way that allows them to deliver the best care for patients. are any answers). I ~ O X + wayr'4~r~yt-a~ ~m iqu&on tfie For example, a major innovation in managing patients with ~ ~ B W E ~ C ~ I U @ : D~w~lw~~~~@~By1h4k~aYre"pio~s'LPBn~f ' " OHTN involves using risk calculators to assess the chance of ~ ~ & & d d a ~ ~ @ ~ m b a w w ' ~ ~ c ~ w ~ 3 ~ ' ~ e tMdmce? developing POAG. km .&Qesd&@~~;.,sm~hp~a--q What is the ideal second-line treatment for POAG? What is @~~&~&~,ab~th~ianGH~phy~~~hahoul&-M the best surgical approach for a patient with pseudophakia ~ ~ ~ ~ ~ ~ & % l ~ ~ ~ ~ ~ ~ @ ~ l r n m m ~ . ~ Mansberger and uncontrolled IOP? Many of these questions will be introduced a glaucoma risk calculator to estimate the proba- addressed in thesubsequent chapters ofthis textbook. Finally, bility of converting from OHTN to POAG in 2003 using data practicing EBM allows you to challenge seemingly authori- from 0HTS.'4 In 2005, Medeiros and colleagues confirnled tative expert opinion. If a learned glaucoma expert comes to the function of a similar calculator in an independent sample your town espausil~g a statement like, "Visual field progres- of patients with OHTN.15 In 2007, researchers from the sion from POAG can be completely halted with the addition United States and Europe introduced a revised calculator of new agent X," then you can use the EBM process to see if based on the combined results of OHTS and the European such a statement is justified.

L.R. Pasquale

2.2.3 The Evolution of Glaucoma Care

Since EBIvI centers on using published evidence to guide clinical decision-making, it is useful to review how ophthal- mic advances influenced our managerial approach to the glaucoma patient. It is difficult to txeat a disease if you do not even know what type of condition it is. It is somewhat ironic that the word glaucoma has its origins in the Greek language and it roughly means "grayish gleam" -hardly a term associ- ated with an optic nerve disease. It originally was a term referring to elderly people with seemingly white and quiet eyes and a dull appearance on external inspection that was associated with visual disability from multiple causes includ- ing mature cataract. Thus, prior to the invention of the oph- thalmoscope by Hemholtz" (before 1551), visual disability from glaucoma was frequently confused with other condi- tions. The postoplzthalmoscope era afforded the clinician an ability to inspect the optic nerve and identify a group of visu- ally disabled people who had white and quiet eyes, relatively clear media, and excavated optic nerve heads. Palpation of the globes of many (but not all) of these patients suggested that the IOP was elevated (the Schiotz tonometer had not been invented yet). These observations established that glau- coma was an optic nerve disease, but the frequent association with a firm globe to palpation suggested that glaucoma was a condition of elevated IOP that produced pathologic optic nerve changes; it is now known that glaucoma can occur across a spectrum of IOP values that include those in the statistically normal range. Nonetheless, the invention of the ophthaln~oscope allowed clinicians to direct IOP-lowering treatment to patients with pathologically cupped optic nerves and elevated IOP. Intuitively it made sense to lower IOP in these patients with the limited therapies available to achieve such an effect. No one dared to question that an RCT might be needed to confirm that patients would benefit from such treatment.

A clear understanding that glaucoma existed in an angle-closure form and an open-angle form was not apparent until after the gonioscope was embraced as an important diagnostic tool. The postgonioscopic era began around 1920 when Koeppe and others perfected the technique of evaluating the filtration apparatus with the patient in the seated position.I8 Ultimately, this technique allowed physi- cians to more accurately identify patients with angle closure glaucoma who would benefit from iridectomy. Interestingly, despite the clear-cut importance of gonioscopy in a glaucoma work-up, available evidence suggests that gonioscopy is relatively underutilized in the management of glaucoma patients t~day.'~.~O

a ~ a ~ i : ~ ~ ~ h & d m ~ ~ ~ u @ ~ W m ~ ~ ~

available had either undesirable side-effect profiles or poor efficacy. Pilocarpine, which has reasonable efficacy and has

been available for more than 130 years to treat g l a ~ c o m a , ~ ' requires frequent application and produces a constricted pupil. In younger patients, these drugs produced significant myopic shift because of induced accornrnodntion of the lens, and in myopic patients, they would occasionally cause retinal breaks.22 Topical epinephrine had a slightly better ocular side-effect profile but limited efficacy. Systemic carbonic anhydrase inhibitors were available from 1954, but they exposed the patient to systemic side effects. Th&.hbzoduotiofi W@Mblsl in1 I19T8 alsharet4 f114he goldeh age WrhC ?op%a.l B.ibBaok&fi Topical beta-blockers were welcomed with open arms because while they did have some systemic side effects, they were tolerable and these drugs provided accep- table efficacy with once-daily or twice-daily dosing. With relatively few published placebo-controlled masked trials using IOP lowering and safety parnmeters as the outcome^,'^ timolol's superiority over epinephrine and pilocarpine became apparent and it instantly became the first-line treat-

-% s r n l q I

ment for glaucoma. 8%~ i r i ' n o one qiit.@uj@~.MethkS J *#,, XI* * x"&Alypr V

?&%6%li'%efa-%Iiockers sfioufd be 6sed Qs"'~ %it'-liae '&gat fo~~1&c?irnmbi~~e4s WE was. a contrai indication m u'sin'g'k 'bkM~'b"lS0aker: *

%?CC jl'99@Vwi@wed #ha polypkrmaqjera, character- ized by a significant expansion of the pharmacological arrna- mentarium in treating glaucoma. The drugs that were introduced (topical carbonic anhydrase inhibitorsz4, alpha agonistsZs, and prostaglandin analogs26) created more treat- ment options for glaucoma patients. In fact, prostaglandin analogs offered the possibility of once-daily dosing, the vir- tual absence of systemic side effects, and IOP-lowering superior to timolol. In the 1990s, the balance had shifted from having limited options for lowering IOP to having a confusing number of ways of dealing with glaucoma, espe- cially when one considers that pamFYaFBdenes ifi l%~Sr"WV- ~ f l ~ E d ~ ' ~ ~ ~ 7 ~ @ l e r 8 ~ t b m ~ 1 ~ & ~ " ~ e a t & e d t'o M wedT&Y.&Efi~4@th~lmtmm~~of glauooma :This time period shares some parallels to the 1940s in general medi- cine when multiple tools were available to address disease (see Fig, 2.1).

A dark cloud was building over the entire field of glau- coma sometime after the golden age of beta-blockers and just before the polypharmacy era when 8pblie p o l + q offi- W~~em~8a!~eE ~$idd J) questioned ~ e ~ e ~ * r n ~ n p Meacal ~ r n ~ ~ ~ r n ~ ~ & w e ~ ~ ( ~ s w t i ~ l w i d &oY Sb&cIficnlly .shgkW%ttt p~il*glaDCm~~ treatmlllenb] qs were ~eal~ly bennefi ting ~ 0 ~ ? 7 The glaucoma field was particularly vulnerable to this criticism because it nmowly focused on IOP as an out- come with precious few studies focused on whether our treatments preserved optic nerve structure and function. This

~f@@%%~e~yg~B@R~~W~eYLTuy~sion~ -

~L$hwa1@9Qsr. Them ~ ~ d ~ h t ! l p ~ ~ k ~ ~ ~ ~ a Y ' l o w e r i n g I W w 1 ~ t ~ ~ m W d e v d l r " ) ~ e ~ ~ o I ~ C t ~ o n g r patients

2 An Evidence-Based Approach t o Glaucoma Care 27

1~4th The OHTS2' and slow disease progression in those with early manifest open-angle glaucoma (the Early Manifest Glaucoma Trial).2Q In order to refine which IOP-lowering modalities were best for our patients, other RCTs cornpared a medicine-first approach to laser-first (the Glaucoma Laser Trial)lo and medicine-first to surgery-first (the Collaborative Initial Glaucoma Treatment Study) in managing open-angle gla~coma.~ ' These trials will be the subject of later chapters, but they can be summarized by stating thar~rwerall them is, currenbly no clear-cut advantage to using one modality to lowar IOP when compared with anether $or gl~auaorna. Nonetheless, I will demonstrate that the EBM process can be used to inform clinicians about the management of individ- ual glaucoma patients.

What will the future hold for glaucoma care and how will EBM help shape it? To answer this question, it is important to hark back to the time before the "postophthalmoscope" era. Initially, we did not even know that glaucoma was a dis- ease of the optic nenre. Today, POAG - the most common foim of glaucoma in the Western world - is a condition that is described in terms of risk factors rather than patho- etiologic elements. Similarly, primary angle closure glaucoma (PACG) - a common form of glaucoma in the Eastern world - is also poorly understood at a fundamental level. Thus, the rfbxt +mm,fn ~ l a ~ ~ ~ m ~ m a n a g e m m ~ t ~1l '~srem from aamore c@mpleTrf! fibdemanding-of lhh~ oombiaation G'Pgeh'&C and

Then, statistical methods are applied to determine which studies from disparate centers can be combined. Models are then formulated to convert a series of slllall trials into a sin- gle lage trial. The advantage of this approach is particularly apparent if some of the smaller trials actually contradict one another, or if the individual trials conclude that one treatment is similar to another, but each individual study is actually underpowered to detect any real difference between the treat- ments. For example, a meta-analysis of nine RCTs concluded that birnatoprost and travoprost are slightly Inore effecti1-e in lowering IOP than latanopro~t.~? This finding is consistent with another meta-analysis of 13 RCTs that concluded that bimatoprost was superior to latanoprost in lowering IOP.33

Nevertheless, more work is needed as still another litera- ture synthesis has pointed out that @o indiddual ~nedicaf 4g,eg.t Bas been di~ovn? to preserve optic nerve hnction as

,&G@@G$~&, aut~m~tetil p&Irie@y, and dnly beta-blockers A

haw bam6,showa to slow dis~ase prog~ession:~~ The absence of such e~~idence does not mean that our current therapies are not effective at presening vision, but more clinical research is needed to optimize our treatments for glauconla.

tnV'ifbirrrYetrMk~fa8to~s thB,diahte 1tha.ktiwlogy of dl hl-rn~ of gl~ucbma.'As the various combinations of genetic determi- 2.3.1 Formulation of a Clinical Problem nants and environmental influences that dictate disease are published and confimled, they w ~ l l become thwbmirsiorgean- Assume a 70-year-old white female presented to your office o t j p ~ @ E e i f i ~ . ~ b & l o r & ~ a $ ~ a ~ s , which may presage the today with difficulty driving at night. Your examination leads ''t~dfi+i'd~?-dZie~b'~i~edi~~rIe ~ i k ' ~ d f h , f ~ ~ c d f f a a a ' EBM will be you to conclude that the patient's symptoms are related to useful to determine if novel genotype-specific strategies to mild cataract formation, but an incidental finding was the treat glaucoma are superior to conventional approaches of presence of exfoliation precipitates in the anterior segment managing the condition. Initially, genotype-specific strate- of both eyes, IOPs of 26 mmHg OD and 21 mmHg OS, and gies may not be embraced (initially, there was resistance to open angles on gonioscopy. Furthermore, the cup-disc ratio accept Schiotz tonometry as a replacement for finger palpa- is 0.8 OD and 0.55 OS. On Humphrey visual field testing, tion of the globe in the assessment of IOP), but RCTs will there is a superior arcuate defect and inferior nasal step in the pressure physicians to adopt these approaches if they are right eye and a shallow superior nasal step in the left eye. The truly superior to the way we practice currently. Snellen acuity is 20130 OU in both eyes. After you explain

~VWC%WM~~)I,~'% Grtlll'sek iribdfflcationk bP&c cuff!nf"%at cataracts were making it difficult to drive at night, the - h ~ ~ ~ m ~ ~ o l ~ ' f i $ P " c ~ r f ~ I d W P %fiflzm vfidgefice" bf a patient opts to defer cataract extraction because she was not

~ % i ~ ~ ~ b ~ 1 d @ w l l a t @ m * & f l ~ p ~ a P e " * ' 2 E ~ fl$?F-~'f"-d9~e$op&g critically disabled by her visual complaint. She was actually @fl~u@nB8f&t?l MP%SP~P'~YO~&SS"?~~~Y~ $;e' ,&;;"i'e'*for-fif relieved to know the major source of her vision probIem was &@d8eas?:*Such knowledge will be based on a more com- potentially fixable. Nonetheless, she was not prepared to dis- plete understanding of the natural history of disease and how cover she had glaucoma. You infornl the patient about the it is modified by current treatment, Such knowledge will, of diagnosis of exfoliation glaucoma and, without much reflec- course, be evidence-based. W %Y.Iv8l$b ~p'"'@'W~~cwmwim tion on the matter, initiate treatment with travoprost 0.0047' ~~.a~e0E.mm-~m1~~~"s't't~~~mesf~c~~~~~IE~u'iii'~1~t- ophthalmic solution dosed one drop at bedtime in both eyes. ~' ' 'eWd~@e~~rn&ing~~rnr * c ~ m e g1'&flt!8fffgg"f$eaflen?s. The patient readily accepts this treatment; but in a quieter Me@-analysis is really a formal statistical approach to review moment, you wonder whether such treatment was the the literature on a subject. First, the quality of the evrdence optimized approach to her case. ~$%8R6m?of~Mee~bf i v is assessed so that only worthwhile reports are included. ~~~&mdw&~~9~BW~p1@~~~brn~ati@fi~tirf a

clinical question.aIn this case, the question is what is the best first-line treatment for this elderly lady with new-onset exfo- liation glaucoma?

2.3.2 Assembling the Evidence Cart: From Textbooks to Systematic Reviews

After crystallizing a clinical problem based on a patient encounter, the next task is assembling the available evidence that addresses this particular issue. Your question really is: "If I am committing this patient to medical therapy, is travo- prost really the best choice for her?" Stated another way: Is there a prostaglandin analog that is particularly effective for exfoliation glauconn? The meta-analyses referenced in the last section indicating that travoprost and bimatoprost were superior to latanoprost did not stratify their results by open- angle glaucoma subtype ( i t . , exfoliation glaucoma versus other types of g l a u c ~ m a ) . ~ ~ . ~ ~ This reformulation of the ques- tion assumes that medical therapy, as opposed to laser trabe- culoplasty or incisional surgery, represents the best initial approach to managing her condition - an issue we will return to later. Furthermore, in thinking about the answer to this question, we must also consider the ocular side-effect profile of each agent.

In general, there are many potential resources available to potentially address your patients' problems, and oftentimes, these sources contain conflicting information. They include textbooks, "blue ribbon panels" convened by cost-conscious third-party payers, consensus panels handpicked by phar- maceutical companies, journal articles, online search engines such as PubMed, or online databases that search multiple articles and synthesize the literature on a particular subject. Where would you go first to find the answer to your question?

Thomas Kuhn in his highly influential treatise entitled, The Stnicture of Scientific Re~o lu t ions ,~~ points out that, "Textbooks aim to communicate the vocabulary and syntax of a contemporary scientific language.'' For example, a typi-

- % i t ~ t b ~ k , iv ~~seih.l.~~tm ~ 1 0 &w ~ b @ & ~ d 'abo~t &@1!@&fli~1~~0@ITf!la RfldY&f d~!3etl!?Ff%8h.$@fh"C2Tlt $fflwwM t ~ m ~ $ , 'BWlT'+$ ~nlWYy LG glftieym om h@psp@@if- ~w~@fi~@pbi~m~lim~.n?~~~~~~~~"t 0Y'~oui . '~0-~ear- 'o~d ' r"ddy '~ ~ ' ~ ~ d ~ ~ ~ b H ~ ~ ~ f ~ f t S i l ' gI$21Cbm~(.*You may seek the assistance of the personal library of journal articles that you subscribe to and read faithfully every month, but it will become obvious very quickly that scouring them will not be a terribly efficient way to find the answer to your question. The logical way to address this problem is to use a general online search engine like PubMed to find studies that address your particular question. In order to use PubMed, your patient encounter, which has been transformed to a relevant

L.R. Pasquale

clinical question, must be transformed again into "keywords" that can be searched, If we had used the keywords therapy and exfoliation glaucoma (other synonyms such pseudoexfo- liation glaucoma should also be tried) on June 21, 2005, there would be 319 references and most of these would not be relevant to this patient's individual clinical situation. In reviewing this list of publications, we find five articles3M0 that are relevant to our clinical problem.

While we were able to find relevant information about our patient problem on PubMed, this is a sornewKat ~ ~ c ~ ~ ~ g , p ~ o q g + t j s . Is there a better way? The short answer is: not at the current time. Theoretically, the Cochrane C~llaboration,~ an organization that provides up-to-date reviews on healthcare interventions in all areas of medicine, would be the ideal shortcut to get the answer to our patient- related question. Essentially, the Cochrane Collaboration searches for articles that perform the kind of literature search outlined above and summarizes the results in one article. At this time, a search using keywords therapy and e~foliation gla~rcoma or ~goliation glazrcoma done did not yield any relevant results sincema ofit! Wmm .an.-evidenoe-based muiaw @nr lehe bhbjeab of aptimum medical'therdpy Ybr &oliahn glaucoma In fact, the keyword glazrcomn will yield a list of 56 meta-analytical articles on glaucoma in all.

2.3.3 Evaluating the Quality of the Evidence

Once the evidence cart for our clinical question has been assem- bled, we must examine the quality of that evidence. Certainly, if the five articles we found related to our patient were of poorly designed studies, then they would not be useful in guiding our decision-making process with respect to our patient. Cochrane reviews would ordinarily do the quality analysis for us, but again there are no such reviews related to our patient. It is recom- mended that when the evidence cart is assembled that one looks for the RCTs first, as they represent a gold standard level of evidence for any paaicular intervention. '&hu~ona n d s r h , rak k$$&&.&&,$~~@8;$,1@b~~ t t l@~ ~c!Jk:

Is the allocation to treatment arms random and, if possi- ble, are the investigators masked to the interventions employed in the study? Are the baseline attributes of each treatment arm clearly stated? A particularly useful study design incorporates a crossover paradigm. In crossover studies, each patient serves as his or her own control and is exposed to each treatment after an adequate washout period. This addresses any difference in baseline attributes between groups and also serves to increase the power of the study. Is the study adequately powered to find a difference between treatment groups? Look for a mention of an

2 An Evidence-Based Approach to Glaucoma Care 29

a wiol'i power calculalion in the methods section of the 2.3.4 Applying the Evidence to Our Case paper. bo the authors present an intention-to-treat (ITT) analysis? There is an important dictum in clinical trials: once randomized, always analyzed. Patients who switch treatment arms or who drop out are always arlalyzed in the groups they were allocated to at the time of randomization. Are withdrawals adequately reported and is [he dropout rate I-easonable (below lo%)? Is there industry sponsorship for the study? Industry sponsorship does not mean the study is tainted but it is important to take note of such sponsorship. It is almost certain that industry sponsorship creates a certain publi- cation bias, whereby results that do not show a product in a favorable light are suppressed, New rules about reg- istering clinical trials with the Food and Drug Administration (FDA) are trying to minimize such pub- lication bias.

Let us look at the evidence we have accrued for our particu- lar clinical question. All five studies are randomized clinical trials. The majority of studies involve Greek nationals, raising some concerns about the generalizability of the data to other populations. A surnmary of the quality of each article is provided in Table 2.1. Each article gets a point each if the masking is judged to be adequate, power calculations are provided, an intent-to-treat analysis is performed, the withdrawal rate is less than 10% (somewhat arbitrary), and industry support is declared. If the study is free from industry sponsorship, then a score of two points is given. More formal grading systems of papers can be performed, but since the purpose is not to perform a meta- analysis but to get a rough idea that the data in the paper are believable, a simplified grading system looking for these particular attributes is recommended. For the evidence, we have accumulated for our question, we have five studies with quality scores ranging from 1 to 6 on a scale of 0 to 6, with most studies having scores of 4 or more. Therefore, the quality of the evidence that we have accumulated is generally quite good.

After assessing the o~erall quality of the evidence assem- bled, it is important to ascertain the outcome of these stud- ies. All of these studies looked at either diurnal IOP or IOP at o l ~ e time point as the main efficacy outcome. There are no data regarding whether any of the treatment options available to treat our particular patient are more effective in preserving vision. In looking at the IOP outcomes, it is also important to distinguisli between statistical significance and clinical sig- nificance. Regulatorp agencies state that differences in IOP of 1.5 mmHg or more constitute clinical significance, although there is no evidence to refute the notion that snialler differences in IOP that are statistically significant may also be clinically significant. The side-effect profile of the various treatment options should also be considered In making a decision about our patient.

In summary, the evidence (summarized in Table 2.2) indi- cates that &or ~ w f ~ b t i o n glavo@mB, Qoth wavoprost and: b&pt~pwst g v ~ d u m mti9@c!aMy significant + lo~irerlng of' dp+mql,J,QP xessus, Latanopmst, but the differences in IOP lowering are less than 1.5 mnlHg. Latanoprost produced sig- nificant reduction in diurnal IOP versus timolol that was greater than 1.5 mmHg. Two studies indicated that fixed co~nbination timolol0.5%-dorzolamide 2% was supeiior to travoprost and latanoprost in terms of IOP lowering, but the difference was 4 . 5 mmHg. Furthermore, one of these stud- ies did not receive a high quality score and the other only used 10 AM IOP as the outcome of interest. From a clinical perspective, it might not be a good idea to start newly diag- nosed patients on fixed combination therapy because it exposes a newly diagnosed patient to the side effects of two classes of medicines. Overall, no major safety issues were raised in any of these studies. Ba@&wdecision To srm the "

ng&@& a but 'bne c0~1d dot be ~ult.eLt&@~a~~~'~any~F;rh~~~oswglrndin analog for our n a w 1 ~ d l f i ~ E d t!X!fblI~doh ,aliid8rnd pat5ehtY

Is it reasonable to consider laser trabeculoplasty or trabeculectomy for our patient? There are no RCTs designed to compare medical therapy to laser trabeculoplasty

Table 2.1 Quality scores for randomized clinical ~ a l s assessing tile effectiveness of medical therapy for exfoliation glaucoma

Masfing ITT analysis Withdrawal Industry sponsorship Quality Study adequate? Study design Power adequate? performed? rate delineated? score

Konstas et a136 Yes Cross-over: N=40 Yes Yes 5% None provided 6 Konstas et aP7 Yes Cross-over; N= 129 Yes Yes 5% Pmvided 5 ~armaksiz et al3s N~ Parallel; N= 50 Not assessed Yes 16% Unclear 1 Konstas et a139 Yes Parallel; N= 103 Yes Yes 5.5% Partial 5 Konstas et alw \'es Cross-over: N=65 Not assessed No 16% None provided 4

m i n t e n t to treat

L.R. Pasauale

Table 2.2 Randomized clinical trials assessing the effectiveness of medical therapy for exfoliaiion glaucoma

Study Quality score Outcorne measured Efficacy Magnitude of effect Konstns et nlZ6 6 Diurnal IOP Travatsn>Latanoorost 4 . 5 mmtIo diserence Konstas et aljl 5 Pam~nksiz et alss 1

Konstas et aP9 5 Konstas et a140 4

- Diurnal IOP Birnntoprost>L3ranoprosta 4 . 5 mmHg difference Diurnal IOP Dortolnmide-Timolol fixed cornbination> 4 . 5 mmHg difference

TravoprosvLatanoprost Diurnal IOP Latanoprost>Timolol 11.5 mmHg difference 10 A ~ I IOP Donolamide-Timolol fixed No

combinationrLatanoprost

"Favors industry-sponsored product

(performed with either the argon or selective laser unit) in with OHTN who were observed converted to a diagnosis of exfoliation glaucoma. Nonetheless, one can refer to the POAG, while only 4.5% of those under treatment converted Glaucoma Laser Trial, which showed that argon laser trabe- to the same diagnosis. The ARR is 9.5-4.5 or 5%. The reln- culoplasty was as effective as medical therapy (during the tive risk reducriorz (RRR) is the proportional reduction in golden age of beta-blockers) in the treatment of open-angle event rates (failure to respond to treatment) between ueat-

Thus, if there are unusual circumstances that ment option 1 and treatment option 2 in a clinical trial. In the prevent adherence or if the patient and the physician simply OHTS, treatment did not completely prevent patients with prefer a laser-first approach after an appropriate informed OHTN from developing glaucoma. In fact, treatment resulted consent process, it is reasonable to proceed with laser trabe- in a 4.519.5 or 5 0 8 RRR. Most diseases have low rates of culoplasty to manage the case. With respect to surgery, the occurrence; therefore, RRR tends to inflate the apparent Collaborative Initid Glaucoma Study evaluated whether benefit of any given treatment and ARR tends to underestimate glaucoma filtration surgery or medicine is the optimal initial it. Thus, another term has been introduced into the calculus approach to manage newly diagnosed OAGS3l The interim of EBM to put the relative effectiveness of treatment into study results showed that visual acuity and visual field out- perspective, and that term is the nzlrnber needed to treat comes were similar in both groups after 4 years of follow-up, (NNT). The NNT is the number of patients who need to be prompting thefmdyinW!$@ga:fdrs to cohclude &LpRyd6ian~~ treated in order for one patient to receive benefit (such as the w h ~ d d mot @hang@ thck practice "ptE!m's' aP rotMakIy .nor E"'prevention of one case of POAG from developing among a o E ~ g a t i e n t s , s l l r g e t ~ d as thei firetoption to manwe OAGg group of patients with OHTN). The NNT is calculated as 11

ARR, rounded to the next highest whole number. For the OHTS, the NNT is 20 - that meme one. w ~ d d

4&,gr&,~wt ~ ~ ~ a t i b r n ~ - w i . ~ OHTN to prevent one of

2.3.5 The Basic Caicuius of Evidence-Based Me dicin e

A typical RCT will report the main outcomes of a shldy and discuss the side effects associated with one intervention versus another. For example, in the OHTS, which assessed whether prophylactic treatment of OHTN prevented people from devel- oping POAG, it was reported that the rate of conversion from OHTN to POAG was reduced from 9.5 to 4.5% with ocular hypotensive therapy after 5 years.2s !hf?bk ?Iaa &g@!3 tllR ~1bw~rin~TOPkn~~OH~~hCIpSi people ieierdly. "by& we @f ' a ~d~ep~~p%~nderslf~nding~of Theytrial ~E.Wlts? Furthermore, how do we use these data to guide treatment decisions for individ- ual glaucoma patients? The terms absollite risk reclziction, relative risk reduction, and the ?lzimber needed to treat4' help us assess exactly how effective ocular hypotensive therapy is in reducing conversion from OHTN to POAG.

The nbsobrre risk redtiction (ARR) is the difference in outcome rate between treatment option 1 and treatment option 2 in a clinical trial. In the OHTS, 9.5% of'patients

ZQAQ a&,&v@l@phgb A lower value for the NNT is desir- able for any treatment intervention and 6WlrTNT M the rM@ M,&iiiis @?mernlly mgtftded h$ fndiaatim of lhn effkttivk &~ammb( the NNT for the Early Manifest Glaucoma Trial where newly diagnosed OAG patients were randomized to treatment versus observation was 6),29 but ahighei'NNT ~ & y b@8m~&bleiifmkp6aphy1aa~c hreatrnen'lS. The OHTS dnh %t oonsidered $a pmpkpl~@tIdlor preventive trial:

Interestingly, if the RRR stays constant, but the number of people reaching an endpoint increases, then the NNT goes down. A case in point regards the subset of African- Americans enrolled in the OHTS. The 5-year risk of convert- ing from OHTN to POAG for African-Americans i n the OHTS was 8.4% in the medical therapy arm and 16.1% in the observation arm."2 Thus, the ARR is 16.1-8.4 or 7.7% and the RRR is 8.411 6.1 or 48%. Overall, the NNT is 1/0.077 or 13. So among African-Americans, one would have to treat 13 patients with OHTN to prevent one case of POAG.

The ARR, RRR, and NNT provide quantitative insights into the results of RCTs yielding a global estimate of treatment effectiveness, but they do not tell us which

2 An Evidence-Based Approach to Glaucoma Care --

individual patient requires treatment. The reader should be cautioned from concluding that all African-Americans with OHTN should be treated because the NNT is 13 when com- pared with an NNT of 20 overall. The decision to treat indi- vidual OHTN patients is better guided by the glaucoma risk calculator. The calculator, which has been validated in patients outside the OHTSI5, accounts for important risk fac- tors that predict coilversion from OHTN to POAG and pro- vides a 5-year risk estimate of converting from OHTN to POAG. The parameters that go into the calculation are age, meail IOP, central corneal thickness, vertical cup-disc ratio, and mean defect on the Humphrey visual field. Race is not entered into the calculator because multivarinble analysis indicated that race was not an independent risk factor for conversion from OHTN to POAG. So a 38-year-old patient with mean IOP of 14 mmHg, OU; CCT of 520 pm OU; CDR=0.4 OU; and PSD of 1.5 dB OD and 1.7 dB OS has a 5-year risk of converting to POAG of 15%. If one uses the calculator to stratify patients into low (<5% in 5 years), medium (5-159'0 risk in 5 years) or high ( > I 5 8 risk in 5 years) risk of converting to POAG, then this person, regardless of race, would have a moderate-to-high increased risk of converting to POAG. As it turns out, African-Americans will tend to fall into the high-risk category by virtue of the fact that they tend to have thinner corneas than their Caucasian counterparts. Nonetheless, an African-American subject with thick corneas does not necessarily have to go on treat- ment because the NNT for African-Americans overall is 13. Ultimately, the risk calculator should not be viewed as tlle instrument that rigidly guides treatment decisions, as that is an individual decision between the patient and doctor.

All treatment interventions have side effects relative to observation without therapy or an alternative treatment. Thus, &j&&$pfy& ip p@$& sbn, adv,wie , G ! X ~ G ~ ~rofd0 in quanti- M & e i p ~ e b C v 6 as wdl hen eonddieritfg Weament for the ' i a & j ~ d u d patiam The Wmwtmbsr nes&d-ro rhm J ~ H ) is the number of patients that need to be treated to harm one patient. In the Collaborative Initial Glaucoma Treatment Study (CIGTS), patients with newly diagnosed OAG were randomized to treatment with medical therapy or surgery to lower IOP. Both treatments were equally effective in stabiliz- ing vision and the visual field over a 5-year period. In fact, both treatments seem to have equivalent impact on quality- of-life measures. Nevertheless, 9% of patients in the medical therapy arm developed cataracts requiring surgical interven- tion, while 20% of patients in the surgical arm required cata- ract extra~tion.~' Thus, the NNH is calculated as 110.2-0.09 or 9. Thus in treating nine newly diagnosed patients with surgery first, the clinician would induce one extra cataract requiring surgical removal. A low NNH is not desirable, especially when the treatment inducing harm is not better b a n the alternative therapy (in this case medicine first). Wit11 that said, the "harm" in this case was a cataract that produces

vision loss that is reversible with cataract extraction. In fact, CIGTS researchers show that cataract surgery in patients with prior filtration surgery is quite s u c ~ e s s f u l ~ ~

While the NNT and NNH provide quantitative interpreta- tion of RCT results in terms of how they benefit patients, they do not tell us if treatment is cost-effective. Cost- efiec&vi\lenassJis a somewhat contro~rersial term because it indicates these is a OOSE above which society cannot bear the burden, of tne~ting a speeifiu condition. We like to think that ideally we should spare no cost in preventing glaucoma, but in reality, resources are limited and must be distributed judi- ciously such that the most treatable conditions are addressed on a societal level. Certainly, we ulould not want to spend so much money on treating OHTN that there are no resources available to care for patients with cataract or progressive POAG. The costs of healthcare have risen by about 7 0 9 per year from 1998 to 2004, and societjp needs to allocate finite resources for the most cost-effecti~e treatments.J4 The National Institute for Clinical Excellence has set a ceiling for annual costs associated with treating nny disease, which reflects the resources available in an affluent count* Basically, if the incremental cost-effectiveness ratio (ICER) is greater than a societal cutoff for willingness to pay, the treatment is regarded as not cost-effective. The upper limit for ICER may be much lower in countries of more limited means. The ICER is calculated using RCT data, the actual costs of treating or observing the condition in question, and accounting for the number of years of life people would be willing to swap in order to remain free of the condition in question. The ICER for treating all OHTNpatients is approx- imately $89,000?5 That means the incremental cost of treat- ing all OHTN patients to prevent one case of POAG is $89,000, a value that is of borderline cost-effectiveness even for a highly developed country like the United States. If we limit treatment, for example, to those patients with CCT 40 pni below the norm (550 pm), then the ICER is close to $37,000. Similarly, it is cost-effective to treat OHTN patients over the age of 76 (two decades above the age of 56), IOP >29 mmHg (4 mmHg above the IOP of 25 mmHg), and with vertical CDR of 0.6 or more (0.2 units above 0.4) because in each case the ICER is less than $50,000.

2.3.6 What Do We Do When There Is Little or no Evidence?

lrou are a busy glaucoma specialist and you w e emotionally shaken because in the past 2 months you have perfomled two combined cataract extractions with trabeculecto~ny proce- dures that resulted in suprachoroidal hemorrhage - one occurred intraoperatively and one occurred postoperatively.

L.R. Pasquafe

In both instances, the patients developed excruciating pain, 2.4 Conclusion and in one instance, the vision went to light perception (LP) and did not improve. You are trying to decide whether there One of the w0nde1-ful things about the study of medicine is are preoperative measures to consider to prevent this from that the lexrning process continues for a lifetime. Adhering everhappcning to you again. The experience has been devas- to EBM is like maintaining your continuing medical educa- tating for the patients and for you. Your question is quite tion. Physicians will benefit from engaging in the EBM simple: What can be done to prevent suprachoroidal hemor- process throughout their career. Jt d o ~ s an5 to ~ a c k rbage from developing? Very quTclclly wu dfscuvkr that there w @ g p~lactitrt &end$ (silch a s the inhocular injectidn of a e no trials designed to answer this question. In fact, it is m h g i o g e n i o agents in the management of neovasbular likely that there will never be a trial to answer this question @ ~ a o ~ ~ ~ f i @ ) , a n d make rational decisions about adopting because s u p r a c h o r o i d d h e m ~ q h ~ e $ i \r~htivelyrazeev~nt new management schema. One shortcoming is bringing a ~ d a trial would have to include a large numbex af patients EBM to the exam lane. Th'gPBM process is not an exercise

order to be adequately powered to determineif a particular th$L G ~ I L be performed when the patient is in ow exam choir; intewention reduced the rate of supraahoroidal-hemorrhage. however, once we have gone through the EBM process, While suprachoroidal hemorrhage is rare, it seems like a it will certainly be absorbed into our practice patterns and common event when it happens to one of your patients. So help the next patients who present with similar problems. you hope to find some evidence that can help you: .Tn this Furthermore, most of the searches for evidence are manual i~6tanoq the ibe~b source of avidencc That ivan p?b*de the ones because ophthdmio rnetn4in31ytic fopic~eVi&ws ark just answerto yourquestion is observational.sindies, Observational b ~ g i n c i g t ~ - a ~ p ~ t i a the Co&me.database, At the current studies may point to risk factors for the development of t i P ; r ~ ~ m w g~auoarnammagementissues eannotberesolved suprachoroidal hemorrhages. Those risk factors may lead with, &e: GBM p1'9cass. The emergence of the electronic you ro adopt some strategies to rninimize the development of medical record, which requires a computer hooked up to the suprachoroidal hemorrhage. Internet in every exam lane, certainly helps us perform

The source of bleeding in suprachoroidal hemorrhages is real-time calculations of risk for developing POAG in OHTN the posterior ciliary arteries that bridge the sclera and enter the patients. In the end, the EBM process is about delivering the choroidal tissue. The literature review indicates that the risk very best care for our patients. factors for suprachoroidal hemorrhage can be divided into systemic and ocular risk factors. Systemic risk factors include older age, hypertension, atherosclerosis, and intraoperative t n c h y ~ a r d i a . ~ ~ - ~ ~ The insight that you gain from this list is that - --- -- --- ---- -- - - -- -- .- .. - - - - patients need to have their cardiovascular status maximized at the time of surgery. If n patient is running a very high Clinical Pearls

blood pressure on the day of surgery, it might be better to defer the operation to another day. Ocular risk factors include EBM is aprocess whereby the best medicl evidence

longer axial length, aphakia, intraoperative vitrectomy, and , in the literature is applied to patients with specific

multiple prior surgeries. In reviewing this list of ocular risk , clinical problems.

factors, you realize that low IOP, lack of lens support, and The EBM process involves four steps: (1) fornlu-

absence of ocular turgor provided by the vitreous gel may lation of a clinical question, (2) assembly of the

tug on the bridging posterior ciliary vessels causing them to evidence that addresses that question, (3) an

rupture. Thus, you decide that in future you will be careful to assessment of the quality of available evidence,

maintain control over the anterior chamber depth intraopera- and (4) application of the evidence to a particular

tively during all anterior segment procedures. You will avoid patient.

outflow procedures in aphakic patients unless they are The absolute risk reduction, relative risk reduction,

absolutely necessary and even then perform them in a and number needed to treat represent quantitative

guarded fashion. Finally, for your trabeculectornies, whether measures to assess efficacy of an intervention in

they are combined with cataract surgery or not, you will use a RCT. The number needed to harm represents a

laser suture lysis and releasable sutures in a judicious fashion I quantitative measure to assess the adverse effects

of an intervention in a randomized clinical trial. avoiding extreme hypotony, which could trigger a postopera- tive suprachoroidal hemorrhage. These maneuvers intuitively Most clinical challenges in glaucoma do not

make sense and while they may not completely prevent a have a directly applicable randomized trial that

suprachoroidal hemorrhage, they may reduce the chance of addresses the problem.

encountering one. - - - - - . - - - -- -- - --,- - . -- .- A . - - -- - - - . - -. - - .-

2 An Evidence-Based Approach to Glaucoma Care

References

1. The reliable source of evidence in health care. In: Julian PT, Higgins, Sally, eds. The Coclrrarie Hartdbook for Sysfetrlntic Rel$ie,~-s qf h7fervetrriorls. Wiley, Interscience, Sussex England; 2008.

1. Bliandari M, Zlowodzki M, Cole PA. From eminence-based practice to evidence-based practice: a paradigm sh~ft. Mi11t7 Med. 2001;87:5 1-54.

3. Gorlnan C. Are doctors just playing hunches? Tiltte. 2007;169(9): 51-54.

4. Gordon MO. Beiser JA, Brandt JD, et al. The ocular hypertension treatment study: baseline factors that predict the onset of primary open-angle glauconla, rlrch Ophrl~alr~tol. 2002; 120:7 14-730. dis- cussion 87-9-830.

5. Leske hlC, \Vu Sl', Hennis .4, Honkanen R, Nemesure B. Risk fac- tors for incident open-angle glaucon~a: the Barbados Eye Studies. Oplrthnb~~olog~. 2008; 1 15:85-93.

6. Sackett DL, Rosenberg \I'M. Gray JA, Haynes RB, Richardson WS. Evidence based medicine: what it is and u'l~nt it isn't. BAU. 1996;312:71-72.

7. Flynn CA, D'Amico F, Snlith G. Should we patch corneal abra- sions? A meta-analysis. J Fntt~ Pmct. 1998;47:264-7-70.

8. Michael JG. Hug D, Dowd hlD. Management of corneal abrasion in children: a randomized clinical trial. Ann Enlog Afed. 2002;40:67-72.

9. Le Sage N, Verreault R. Rochette L. Efficacy of eye patching for traumatic corneal abrasions: a controlled clinical trial. Anrr Etlrcrg Med. 2001 :3S:179-133.

10. Daly \ I r J , Brater DC. Medieval contributions to the search for truth in cltnical medicine. Persjrect Biol Med. 2000;43:530-540.

11. North R. Benjamin Rush, hlD: assassin or beloved healer? Pmc (Bay1 Urtiv Afed Cerlr). 2000;13:45-49.

12. Bean RB. Sir W~lliant Osler; Aphorisnls. Springfield: Blackwell; 1961:164

13. Morris JN. Recalling the miracle that was penicillin: two memora- ble patients. J R Soc Med. 2004;97: 189-1 90.

14. Mansberger SL. A risk calculator to determine the probability of glaucoma. J Gla~iconm. 3003;13:345-347.

15. hledeiros FA, Weinreb RN, Sample PA, et al. Validation of a pre- dictive model to estimate the risk of conversion from ocular hypertension to glaucoma. Atrlt Ophtl~olrt~ol. 2005; 123: 135 1-1 360.

16. Gordon MO, Torri \I, Miglior S , et al. Validated prediction model for the development of primary open-angle glaucoma in individuals with ocular hypertension. Ol,hthnln~ology. 3007; 114: 10-19.

17. \,on Helmholtz HLF. Bescltreibu17g eines Augeta-Spiegels. Berlin, Germany: A Forstner'sche Verlagsbuchhandlung; 185 1.

18. Dellaporta A. Historical notes on gonioscopy. S~cnl Ophthaln~ol. 1975:20: 137-149.

19. Hertzog LH. Albrecht KG, LaBree L, Lee PP. Glaucoma care and conformance with preferred practice patterns. Examination of the plivate, con~munity-based ophthalmologist. Ophthaltnology. 1996; 103:1009-1013.

20. Coleman AL, Yu F, Evans SJ. Use of gonioscopy in medicare benefi- ciaries before glaucoma surgery. J Glaucoma. 2006;15:486193.

21. Kronfeld P. Eserine and pilocarpine: our 100-year-old allies. Sutv OphAnintol. 1970;14:479.

22. Beasley H, Fraunfelder FT. Retinal detachments and topical ocular miotics. Ophtl~aln~olog~~. 1979;86:95-98.

23. Zimmerman TJ, Kass MA, Ydblonski ME, Becker B. Ti~nolol maleate: efficacy and safety. Arch Ophthaln~ol. 1979:97:656-658.

24. Lippa EA, Carlson LE, Ehinger B , et al. Dose response and duration of action of dorzolamide, a topical carbonic anhydrase inhibitor. Arch Ophthnlmol. 1992;110:495-499.

2% Derick RJ, Robin AL, Walters TR. et al. Brimonidine tartrate: a one-month dose response study, O~~htAaltnology. 1997;104: 131-136.

76. Camras CB, Scl~umer RA, hlarsk A, et al. Intraocular pressure reduction with PhXA31, a new prostaglandin analogue, in patients with ocular hypertension. Arc11 O ~ ~ l ~ r l r a h ~ ~ o l . 1991; 110: 1733-1738.

27. Eddy Dhl, Billings J. The quality of medical evidence: inlplications for quality of care. Health rlff(Afilh~.ood). 19SS:7: 19-32.

28. Kass MA, Heuer DK, Higginbotham EJ. el al. The ocular hyperten- sion treatment study: a randomized trial determines that topical ocular hypotensive medication delays or prevents the onset of pri- maql open-angle glaucoma. Arch Ophrhul11101.1-002;110:701-713. discussion 829-830.

29. Heijl A, Leske hlC, Bengtsson B, Hgman L, Benglsson B, Hussein M. Reduction of inunocular pressure and glaucoma progression: results from the Early Manifest Glaucoma Trial. .4rc11 Ol~l~rl~abrrol. 3002;120: 1368-3 179.

30. The Glaucolna Laser Trial (GLT). 2. Results of argon laser tlabecu- loplasty versus topical medicines. The Glaucoma Laser Trial Research Group. Ol)hrl~nbr~olo~,~ 1990;97: 1403-1413

31. Lichter PR, hlusch DC, Gillespie B\V, et al. Interim clinical outcomes in the Collabomtive Initial Glaucoma Treatment Study comparing initial treatment randomized to nledications or surgery. Ophthol~tlology. 2001;108:1913-1953.

32. Denis P, Lafuma A, Khoshnood B, Mimaud 1: Berdeaux G. A meta-analysis of topical prostaglandin analogues inua-ocular pressure lowering in glaucoma therapy. Crrn' i2Ied Res Opin. 1007;33:601-608.

33. Cheng JW, IVei RL Meta-analysis of 13 randomized controlled tri- als comparing bimatoprost \vith latanoprost inpatients with elevated intraocular pressure. Clin Tl~er. 2008;30;622-632.

34. Vnss C, Him C, Sycha T, Findl 0, Bauer P, and Schmetterer L. ~ W @ @ W ~ I W P J W F ' p ~ r l l p & D m angle. g h u ~ o m a awil Q ~ W h*pmm8io& Cochratle Database Syst Rev 2007: CD003167

35. Ruhn TS. The structure of scientific rei~ol~iations. 3rd en. Chicago: The University of Chicago Press; 1996:213.

36. Konstas AG, Kozobolis VP, Katsimpris IE, et al. Efficacy and safety of latanoprost versus travoprost in exfoliadve glaucoma patients. Ophthalmology. 2007;114:653-657.

37. Konstas AG, Hollo G, Irkec M, et al. Diurnal IOP control with bimatoprost versus latanoprost in exfoliative glaucoma: a crossover, observer-masked. hree-centre study. Br J Ophthalrnol. 2007;91: 757-760.

38. Parmaksiz S, Yuuksel N, Karabas VL, Ozlran B, Demirci G, Caglar Y. A colnparison of tnvoprost, latanoprost, and the fixed combina- tion of dolzolamide and timolol in patients with pseudoexfoliation glaucoma. E~r r J Ophthalrnol. 1006;16:73-80.

39. Konstas AG, hlylopoulos N, Karabatsas CH, et J. Diurnal intraoc- ular pressure reduction with latanoprost 0.00570 compared to timolol rnaleate 0.510 as monotherapy in subjects with exfoliation glaucoma. f ie 2004:18:893-899

40. Konstas AG. Kozobolis VP, Tersis I, Leech J, Stewart WC. The efficacy and safety of the timolol/dorzolamide fixed combi- nation vs latanoprost in exfoliat~on glaucoma. Eye. 2003; 17:4136.

41. Barratt A, Wyer PC, Hatala R, et al. TIPS for learners of evidence- based medicine: 1. Relative risk reduction, absolute risk reduction and number needed to treat. CMAJ. 2004;171:353-358.

42. Higginbatham EJ, Gordon MO, Beiser JA, et al. The Ocular H}~ertension Treatment Study: topical medication delays or prevents primary open-angle glaucoma in African American individuals. Arch Ophthalmol. 2004;122:813-620.

43. Musch DC, Gillespie BW, Niziol LM, et al. Cataract extraction in the collaborative initial glaucoma treatment study: incidence, risk factors, and the effect of cataract progression and extraction on clinical and quality-of-life outcomes. Arch Opkthaln~ol. 2006; 124: 169+1700.

L.R. Passuale

44. Pasquale LR, Dolgitser Ivl. Wentzloff JN, et al. Health care chxses for patients with ocular hypertension or primary open-angle glau- coma. Ophrhnlmralogy. 2008; 115(4):633-638.

45. Kymes SWl, Kass MA. Anderson DR. Miller JP, Gordon MO. Management of ocular hypertension: a cost-effectiveness approach from the Ocular Hypertension Treatment Study. ~4111 J Ophrhr~ln~ol. 2006;141:997-1008.

46. Speaker MG, Guerriero PN. Met M, Coad CT, 13erger A, blarmor M. A case-control study of risk factors for intraoperative supmchor- oidal expulsive I~emorrhage. Ophrhalnzology. 1991;95:201-209. discussion 210.

- - -

47. Obuchowska I, blariak Z. Risk factors of massive suprachoroidal hemorrl~age during extrdcapsular cataract extraction surgery. E u r l Ophrhalmol. 2005; 15:7 12-717.

48. Mosllfeghi DM, Kim BY, Kaiser PK, Sears JE, Smith SD. Appositional supmchoroidal hemorrhage: a case-control study. A111 J Ophihalmol. 2001; 138:959-963.

49. Ehlers JP, Spirn kIJ, Lam A, Sivalingam A, SamueI ?VIA. Tasman W. combination intravitreal bevacizumab/panretinal photocoagulation versus panretinal photocoagulation alone in the treatment of neovascular glaucoma. Rrrinn. 1,00S;3S: 696-702.

Paul N. Schacknow John R. Samples Editors

The Glaucoma Book

A Practical, Evidence-Based Approach to Pat ien t Care

QI - Springer

TAHLEQUAH, OK 74464