Surgical Removal vs Observation for Subfoveal Choroidal Neovascularization,Either Associated With the Ocular Histoplasmosis Syndrome or Idiopathic: Title and subTitle BreakI. Ophthalmic Findings From a Randomized Clinical Trial: SubmacularSurgery Trials (SST) Group H Trial: SST Report No. 9 FREE

Choroidal neovascularization (CNV) particularly when it develops ina subfoveal location, often has a devastating effect on visual acuity andother aspects of vision in the affected eye. In the United States, CNV isdiagnosed most often in eyes with age-related macular degeneration, the ocularhistoplasmosis syndrome, or in the absence of any apparent underlying condition(idiopathic CNV). Neovascular ocular histoplasmosis and idiopathicCNV typically develop in individuals who are in their 30s, 40s, or 50s,¹ often critical periods in their working lives. Unlikeage-related macular degeneration in which CNV frequently develops in the secondeye within a few years of onset in the first eye,^{2 - 3} neovascularocular histoplasmosis^{4 - 5} and idiopathicCNV often remain unilateral. Nevertheless, the effect of this condition onhealth- and vision-targeted quality of life has been documented to be substantial.⁶

Neovascular age-related macular degeneration and idiopathic CNV arediagnosed by ophthalmologists throughout the developed regions of the world,but neovascular ocular histoplasmosis rarely is diagnosed outside North America.In the United States, exposure to Histoplasma capsulatum, which is believed to be the initiating event,^{7 - 9} occursprimarily in the Mississippi and Ohio River valleys and surrounding areas.¹⁰ Although predisposing ocular lesions (“histospots”) develop in only a small percentage of individuals followinginfection,¹¹ there is no known method of preventingeither the predisposing lesions or CNV that may develop decades after theinitial infection. Despite the proven benefit of laser photocoagulation ofextrafoveal and juxtafoveal CNV,^{12 - 18} manypatients are seen initially with subfoveal CNV¹⁹ orhave subfoveal recurrence following laser photocoagulation.^{20 - 21}

Since 1997, the Submacular Surgery Trials (SST) Research Group has beenconducting randomized trials of surgical removal of subfoveal choroidal neovascularlesions to evaluate the role of this procedure in the treatment of patientswith age-related macular degeneration, ocular histoplasmosis, and idiopathicCNV. Patients were enrolled in a pilot study from November 23, 1993, throughMarch 31, 1997. The goal of the pilot study was to refine the design and methodsto be employed in a formal evaluation of submacular surgery.²² Thefirst randomized clinical trial initiated by the SST Research Group was forpatients who had either new or recurrent subfoveal CNV that was idiopathicor associated with the ocular histoplasmosis syndrome (SST Group H Trial).This group of patients was judged to be the one in which submacular surgeryheld the most promise of providing a benefit with respect to future vision,based on reports from uncontrolled case series.^{23 - 25} Moreover,no other treatment had been documented to benefit these patients. Althoughthe benefits of laser photocoagulation for treating extrafoveal and juxtafovealCNV arising from the same conditions were demonstrated by the Macular PhotocoagulationStudy Group,^{12 - 18} investigatorswho participated in a small pilot study reported that there was no benefitto laser photocoagulation of subfoveal CNV secondary to ocular histoplasmosis.¹⁹ More recently, photodynamic therapy with verteporfinhas been suggested to be beneficial for treating subfoveal CNV due to ocularhistoplasmosis²⁶ and subfoveal idiopathic CNV,²⁷ but that conclusion is based on small uncontrolledcase series.

To our knowledge, the SST Group H Trial is the first randomized comparisonof any treatment for subfoveal CNV in eyes with ocular histoplasmosis or similaridiopathic lesions to the natural history of such lesions in a sufficientnumber of patients followed up for a sufficiently long period to provide statisticallymeaningful data. The purpose of this report is to present visual acuity andrelated ophthalmic findings from the clinical trial; a separate report inthis issue of the ARCHIVES²⁸ presentsfindings regarding vision-targeted and general health-related quality of life.The findings from other trials conducted by the SST Research Group for patientswith subfoveal CNV secondary to age-related macular degeneration also arereported elsewhere.^{29 - 32}

The members of an independent Data and Safety Monitoring Committee,appointed in 1996 by the Director of the National Eye Institute, NationalInstitutes of Health, US Department of Health and Human Services, Bethesda,Md, reviewed findings from the pilot study for similar patients and approvedthe design and methods of the SST Group H Trial in January 1997 before enrollmentof patients was initiated. In addition, institutional review boards at allparticipating institutions reviewed and approved the study design and theconsent forms to be used locally. All patients gave signed consent beforeenrollment and random treatment assignment.

The SST Manual of Procedures³³ andthe SST Forms Book³⁴ providedetailed information regarding study design, methods, and policies. Only thoseaspects pertinent to this report are summarized here.

Eligible patients were identified from referrals to 21 clinical centersthat participated in the SST Group H Trial, including 14 clinical centersthat also participated in the SST pilot study. Participating ophthalmologistsand other personnel at each center were required to meet standard criteriato be SST certified to enroll, treat, examine, and follow up study patients.

After a diagnosis of subfoveal CNV had been made by the study ophthalmologist,best-corrected visual acuity was measured following a standard protocol withmodified Bailey-Lovie charts (ETDRS [Early Treatment Diabetic RetinopathyStudy] charts),³⁵ contrast threshold was measuredusing Pelli-Robson charts at 0.5 m,^{36 - 37} andreading speed with enlarged text was measured using charts and methods developedfor the Macular Photocoagulation Study.³⁸ Foreach measurement, the 2 eyes were tested separately; a different chart wasused to test each eye. Stereoscopic color photographs of the disc and maculaof each eye and a stereoscopic film-based fluorescein angiogram were takenby an SST-certified photographer, who followed a standard protocol, to documenteligibility and baseline status for comparison with photographs taken duringfollow-up examinations. Usage of aspirin and anticoagulants and a historyof diabetes mellitus, hypertension, and, beginning in February 1999, othermajor medical conditions were elicited from the patient as part of baselinedata collection.

To be eligible for enrollment and random treatment assignment in thetrial, the patient had to have a subfoveal neovascular lesion that includedclassic CNV. The lesion could be either new CNV (no prior treatment of CNV)or recurrent subfoveal CNV following earlier laser photocoagulation for extrafovealor juxtafoveal CNV as long as CNV was the subfoveal component of the lesion.Earlier treatment of the macula of the study eye with photodynamic therapyrendered the patient ineligible. The total area occupied by the lesion, includingany earlier contiguous laser photocoagulation, could not exceed 9.0 disc areas(DA) (approximately 22.9 mm² on the retina). Eligible eyes hadbest-corrected visual acuity scores of 82 to 18 inclusive, corresponding toSnellen fractions of 20/50 to 20/800 (5/200); visual acuity of the nonstudyeye had to be light perception or better. Patients who had evidence of age-relatedmacular degeneration or other progressive ocular disease in either eye thatcould affect visual acuity or assessment of other ophthalmic outcomes wereineligible. Only 1 eye of each patient (study eye) was eligible for enrollment.Whenever both eyes of a patient were eligible, the study eye was selectedby the patient and ophthalmologist prior to enrollment; the fellow eye wasmanaged as they decided.

Patients judged to be eligible by an SST-certified ophthalmologist weregiven information about the clinical trial and invited to participate. Forthose who agreed, a study identification number and an alphabetic code wereassigned before they completed an interview by telephone with personnel atthe SST Coordinating Center, Baltimore, Md.⁶ Afterthe patient signed the consent form for the trial, the identification numberand alphabetic code were recorded on the baseline data forms that were telecopiedto the coordinating center for preliminary review of eligibility and completenessof baseline data recording. The next treatment arm assignment was selectedautomatically from the electronic file of random assignments that had beenprepared for that clinical center specifically for the SST Group H Trial.The clinical center staff were notified immediately of the assigned treatment,surgery or observation, by means of an automated message returned to themby telecopier. The assignment was communicated to the patient by the enrollingophthalmologist; surgery was scheduled as soon as possible within the next8 days for each patient assigned to that treatment arm.

After enrollment, baseline photographs were sent to the SST PhotographReading Center, Baltimore, for review by trained readers who were masked tothe treatment arm to which the study eye had been assigned. The focus of thereview was on eligibility for enrollment and documentation of characteristicsof the eye and lesion. Lesion characteristics were described by applying standarddefinitions³⁹ adapted from other clinical trials.^{40 - 41} The underlying cause of the neovascularlesion in the study eye was assigned based on the appearance of both eyesand the presence of other lesions in the 2 eyes. The size of the subfoveallesion was categorized using a transparent overlay, adapted from one usedin the Macular Photocoagulation Study,⁴⁰ thathad printed circles with areas equivalent to 2.0, 3.5, 6.0, 9.0, 12.0, and16.0 DA. The areas of the circles on the retina were 5.1 to 40.6 mm², assuming a retinal image magnification factor of 2.5.^{39 ,41} Afinal judgment regarding eligibility was made after review of baseline photographsand evaluation of pertinent baseline data had been completed.

At baseline and each follow-up examination, visual acuity was measuredby an SST-certified vision examiner. Each eye was tested separately afteran SST protocol refraction to obtain the best correction. Testing began at2 m; whenever the patient could read 15 letters or more correctly using theeye, the visual acuity score was recorded as the number of letters read correctlyplus 30. When the patient could not read at least 15 letters at 2 m with theeye being tested, visual acuity was measured at 0.5 m after the refractivecorrection was adjusted for the closer distance. In the latter situation,the visual acuity score was calculated to be the sum of the number of lettersread correctly at 2 m and the smaller of 30 or the number of letters readcorrectly at 0.5 m. Visual acuities from 20/20 to 20/1600 (5/400), inclusive,could be measured using these 2 test distances. Eyes with which patients couldnot read any letters correctly at either test distance (visual acuity <5/400)were tested for light perception with confirmation by the ophthalmologist.

A traveling vision examiner based at the SST Chairman's Office,Baltimore, refracted both eyes and measured best-corrected visual acuity,contrast threshold, and reading speed of patients in both treatment arms atthe 24-month examination, designated the primary outcome assessment examination,and at the 48-month examination for patients enrolled by September 1999. Beginningin June 2002, a traveling vision examiner also conducted masked measurementsof vision at 36-month examinations whenever possible for patients enrolledby June 1999 who had not completed that examination earlier. Traveling visionexaminers were masked to the clinical trial in which the patient was enrolled,the study eye, and the treatment assigned or received.

Only surgeons who met defined conditions based on training and experiencewere approved to perform submacular surgery on study eyes of SST patients.³³ Whenever surgery could not be scheduled within 8days after enrollment, a fluorescein angiogram taken no more than 8 days beforesurgery was required to document the preoperative status of the eye and lesion.

The surgery protocol specified a standard 3-port pars plana vitrectomyand surgical separation of the posterior hyaloid from the macular region whennot already detached preoperatively. The retinotomy site was chosen by thesurgeon to provide optimal access to the neovascular lesion and to minimizesurgical trauma to the fovea. Surgeons had the option of infusing balancedsaline solution to separate the neurosensory retina from the neovascular lesion.Fibrovascular tissue was removed manually with a micro forceps; intraocularpressure was elevated immediately prior to removal of the tissue to minimizehemorrhage from the choroid and was returned slowly to normal pressure afterremoval of tissue from beneath the retina but before its removal from theeye. The peripheral retina was examined for tears prior to fluid-air exchange.A complete fluid-air exchange was performed; fluid was reinstilled to createan air bubble of up to 50% in phakic eyes or up to 100% in pseudophakic eyes.Patients were instructed to remain in a face-down position overnight aftersurgery and until the air bubble was 20% or less. Details of the surgery,deviations from the surgical protocol, and intraoperative complications wererecorded by the surgeon on a standard form.³⁴

Eyes treated surgically were examined as often as judged necessary bythe surgeon during the postoperative period. Patients were scheduled for apostsurgical examination and data collection at 1 month after enrollment.Patients assigned to observation were scheduled for a telephone review ofocular status and history 1 month after enrollment.

All patients were scheduled for follow-up examinations for data collectionpurposes at 3, 6, 12, and 24 months after enrollment. Patients who enrolledby September 30, 2000, were scheduled for a 36-month examination; those whoenrolled by September 30, 1999, also were scheduled for a 48-month examination.In addition to the examinations prescribed by the study protocol, patientswere seen by SST ophthalmologists as often as necessary to evaluate symptomsand the need for treatment of the study eye or fellow eye. For patients notexamined at the midpoint between scheduled SST examinations, interim ophthalmicand medical histories were collected by telephone by local SST clinical personnel.The last study examinations for all patients who had not completed a 48-monthexamination already were performed during the final year of patient follow-upthat began on October 1, 2002.

At the 3-month and later examinations, best-corrected visual acuity,contrast threshold, and reading speed were measured. A detailed history ofsystemic and ocular complications and intervening treatments to the studyeye was elicited. Color photography and fluorescein angiography were repeated;photographs were sent to the SST Photograph Reading Center for interpretationand coding. The emphasis of central review of photographs taken during follow-upwas on identification of CNV and its location in eyes following surgery, measurementof the current size of the subfoveal lesion, and documentation of retinalcomplications. Eyes in the surgery arm were considered for additional treatmentwhenever dye leakage from CNV was observed by the ophthalmologist on the fluoresceinangiogram. The re-treatment protocol specified laser photocoagulation or surgerydepending on the location of the CNV; however, no more than 1 repeat surgerywas recommended. Regardless of the assigned treatment arm, the status of thestudy eye, or the vision of the fellow eye, the study protocol recommendedthat cataract surgery be considered whenever the ophthalmologist judged thatthe severity and location of any lenticular opacity was sufficient to causea loss of at least 2 lines of visual acuity in an otherwise healthy phakiceye.

Prespecified adverse events were reported by ophthalmologists on thebasis of findings at scheduled study examinations; additional adverse eventswere reported as clinical center personnel became aware of them. An AdverseEvent Review Committee, appointed by the Data and Safety Monitoring Committee,classified events and requested additional information when needed to judgewhether the event was in any way related to surgery in the study eye or participationin the clinical trial.

The SST Planning Committee defined the primary successful outcome ofinterest a priori to be improvement or stabilization of visual acuity at the24-month examination, with stabilization defined as a visual acuity no morethan 7 letters better or worse than at baseline. Examinations at 36 monthsand 48 months after enrollment were intended to monitor whether any treatmenteffect observed at the 24-month examination persisted. Sample size was calculatedduring the planning phase under the assumption that 50% of eyes assigned toobservation would have a successful outcome, based on unpublished data fromcases assigned to the observation arm in the SST Group H pilot trial and untreatedeyes of patients who participated in a pilot trial of laser photocoagulationin similar eyes.¹⁹ The minimum clinically meaningfulrelative improvement in success rate after surgery, considering the cost andinconvenience of surgery, was judged by the members of the Planning Committeeto be 50%, that is, a successful outcome in 75% or more of surgically treatedeyes. Available data from 12-month examinations of eyes randomly assignedto surgery or observation in the SST Group H pilot trial and followed throughOctober 31, 1996, suggested that a treatment effect of this magnitude wasrealistic. Type I (α) and type II (β) errors were set at.01 (2-sided)and.10, respectively, to minimize the chance of reaching an incorrect conclusionregarding the effectiveness of surgery during interim monitoring of outcomedata. Allowances for deaths and missed 24-month examinations yielded a targetsample size of 240 patients (240 eyes).

Responsibility for monitoring accumulating data regarding safety andeffectiveness of surgery was entrusted to the Data and Safety Monitoring Committeeappointed in 1996. Informal statistical monitoring guidelines of the Lan andDeMets type⁴² were developed for review atthe meeting of the Data and Safety Monitoring Committee on August 25, 1997;these were based on the assumption that there would be 5 interim analysesof the data on which a recommendation would be made to continue or to haltaccrual. Reviews of accumulating data by the Data and Safety Monitoring Committeeemphasized comparison of distributions of visual acuity and changes from baselineto each examination by treatment arm as well as proportions of eyes in eacharm with successful outcomes at the 24-month examination, the primary outcomedefined in the design.

The Data and Safety Monitoring Committee met in person approximatelyevery 12 months. At the beginning of these meetings, the chair of the AdverseEvent Review Committee reported findings from the most recent interim reviewof potential adverse events to the Data and Safety Monitoring Committee. Attheir meeting on August 22, 2001, with 217 patients enrolled in the SST GroupH Trial, the Data and Safety Monitoring Committee recommended that accrualhalt on September 30, 2001, and that the last patients enrolled be followedup for only 3 years. The rationale for the recommendation was the longer thanexpected period of patient accrual, the slow rate of accrual during the previous12 months, and stochastic curtailment analyses that indicated that there shouldbe little loss of power with the achieved sample size. Subsequently, owingto funding constraints, the follow-up phase of the trial was curtailed furtherso that a minimum of 2 years of follow-up was provided for each patient. OnFebruary 20, 2004, the Data and Safety Monitoring Committee reviewed findingssummarized from the final SST Group H Trial database and recommendations regardingsubmacular surgery.

Data from masked vision examinations by traveling vision examiners wereused in analyses whenever available; otherwise, measurements by local examinerswere analyzed. For analysis of change in visual acuity, eyes with visual acuityworse than 20/1600 (<5/400) but with at least light perception were codedas having visual acuity 3 lines (15 letters) worse than 20/1600, that is,equivalent to 3 lines worse than the smallest line measurable at 0.5 m. Visualacuity scores were used in all analyses. Scores were converted to equivalentSnellen fractions for presentation in tables and text.

The χ² test was used to compare the proportions of eyeswith successful outcomes at the 24-month examination, as defined earlier,and to compare categorical distributions of size and changes in size of subfoveallesions (test for trend). Distributions of visual acuity, changes in visualacuity from baseline to follow-up examinations, and other continuous measurementswere compared between treatment arms using the Wilcoxon rank sum test.⁴³ Proportions of eyes in which visual acuity had improvedor stabilized and the proportion of eyes with fluorescein leakage from CNVat each examination were calculated using a 2-state stochastic model thatconsiders both events and recoveries and compares treatment arms using χ² tests.⁴⁴ Time to first observationof recurrent CNV in the surgery arm and time to cataract surgery and firstreport of visually significant cataract were analyzed using the product-limitmethod.⁴⁵ Eyes were withdrawn from analysisafter the last completed examination when either of these 2 methods was used.Confidence intervals (CIs) on “success ratios” (inverse of riskratios) for surgery vs observation were calculated using the method of Katzet al.⁴⁶

Subgroups of study eyes defined by prespecified baseline characteristicsof patients, eyes, and lesions were examined for consistency of visual acuityfindings by treatment arm. Characteristics to be evaluated were selected primarilyfrom those identified in earlier clinical trials of laser photocoagulationfor subfoveal CNV.^{38 ,47} Of particularinterest were subgroups defined by baseline visual acuity because of findingsfrom retrospective analysis of postsurgery visual acuity outcomes in caseseries and from prospective data for similar patients who participated inthe pilot study that eyes with better visual acuity at baseline sustainedlarge losses after surgery (B.S.H. and N.M.B., unpublished data, February24, 1996). Interactions between surgery and baseline covariates suggestedby subgroup analyses were evaluated using interaction terms in logistic regressionmodels with a successful outcome, as defined earlier, at the 24-month examinationas the dependent variable.

All pertinent data received at the SST Coordinating Center by October31, 2003, were analyzed. Data from each patient were analyzed with the treatmentarm to which he or she was assigned randomly at time of enrollment (“intent-to-treat”analysis approach). SAS (SAS Inc, Cary, NC) and custom-written software wereused for data analyses. P values were not adjustedfor multiplicity of outcomes or comparisons. Although P values suggestive of a difference between treatment arms (P≤.05 for overall findings and P≤.10for prespecified subgroups) have been noted in this report, only P≤.01 was deemed statistically significant for overall findings,as specified in the design.

From April 1, 1997, until accrual ended on September 30, 2001, two hundredtwenty-five patients (225 eyes) enrolled in the SST Group H Trial; 112 patientswere assigned to the surgery arm and 113 patients to the observation arm.Based on central review of baseline photographs, 27 eyes, 17 (15%) in thesurgery arm and 10 (9%) in the observation arm, were judged not to meet thestrict criteria for eligibility. In 14 eyes (10 in the surgery arm and 4 inthe observation arm), CNV was judged to be in a juxtafoveal rather than subfoveallocation; 2 of these eyes, both in the surgery arm, also were judged to beineligible for other reasons. Other reasons for judging eyes to be ineligiblewere evidence of some other cause of CNV (6 eyes, 3 in each arm; 2 eyes withanother reason also), no classic CNV (5 eyes, 1 with another reason also),total size of the subfoveal lesion greater than 9.0 DA (1 eye in each arm),and photographic quality inadequate to judge eligibility (1 eye). In addition,1 eye in the observation arm that also was ineligible for other reasons hadbaseline visual acuity worse than the Snellen equivalent of 20/800 (5/200).Data for ineligible patients were analyzed with all other patients in thetreatment arm to which they were assigned.

Characteristics of patients, eyes, and subfoveal lesions at the timeof enrollment are summarized by treatment arm in Table 1 and Table 2. Treatmentarms were well balanced on all characteristics recorded. The median age ofpatients was 48 years; 57 patients (25%) were 60 years or older. Almost allpatients (96%) classified their race/ethnicity as non-Hispanic white, consistentwith epidemiological data and patient populations enrolled in other clinicaltrials of treatment for idiopathic CNV and neovascular ocular histoplasmosis.^{12 - 15} Asa group, the remaining patients were distributed equally between the treatmentarms; however, 4 of 5 non-Hispanic black enrollees were assigned to the surgeryarm. Seven of 13 patients who reported that they were unable to work attributedthis inability to poor vision. Information about other medical conditionsat baseline has been published.⁶

Median visual acuity score of study eyes at baseline was 66 letters(20/100 Snellen equivalent). Median reading speed with the study eye alonewas 66 words per minute; median contrast threshold was 4.5%. The lesions instudy eyes were judged centrally to be due to ocular histoplasmosis for 192patients (85%); 64 study eyes (28%) had subfoveal CNV that had recurred afterlaser photocoagulation. Almost half the subfoveal neovascular lesions in thestudy eyes were 2.0 DA (5.1 mm²) or smaller in size; only 20 eyes(9%) had subfoveal lesions larger than 6.0 DA (15.3 mm²). Medianvisual acuity score of fellow eyes at baseline was 100 (20/20 Snellen equivalent).Fifty-six (29%) of the 192 patients with neovascular ocular histoplasmosislesions also had a neovascular lesion in the fellow eye; 102 fellow eyes (53%)had other lesions of ocular histoplasmosis but not a neovascular lesion. Two(6%) of the 27 patients with idiopathic CNV in the study eye had a neovascularlesion in the fellow eye.

No patient assigned to the observation arm had submacular surgery inthe study eye during the follow-up period. Of 112 patients assigned to surgery,1 patient refused surgery after enrollment. Two other patients had surgery9 days after enrollment and one 17 days later. In 9 phakic eyes, the air fillat the end of surgery was greater than 50%; gas (not air) fill was used in10 eyes.

In 14 eyes, a small peripheral retinal tear was noted during surgery;12 eyes with peripheral tears were treated with cryotherapy or laser photocoagulation.Two retinal tears in the posterior pole were reported, 1 each in foveal andextrafoveal locations. In 75 eyes (68%) some blood remained after surgery;however, in only 13 eyes was the area occupied by the residual blood as largeas 1 DA as judged by the surgeon.

Four patients in the observation arm died after enrollment; 2 deathsoccurred before the 24-month examination. One patient in the surgery arm diednearly 4 years after enrollment. The number of patients expected to be examinedat each scheduled time after enrollment, based on date of enrollment and vitalstatus, is given in Table 3. Of 225patients who enrolled, 194 (86%) were eligible for a 36-month examination;122 patients (54%) were eligible for a 48-month examination. Of 608 examinationsscheduled for patients in the surgery arm at 3 months or longer after enrollment,570 (94%) were completed; more than 90% of the patients in the surgery armwere examined at each scheduled time. Patients in the observation arm returnedfor 527 (87%) of 606 scheduled examinations. Follow-up examination rates wereconsistently lower in the observation arm than in the surgery arm at all scheduledtimes. Five of 13 patients in the observation arm and 4 of 9 patients in thesurgery arm who were not examined 24 months after enrollment had either a36- or a 48-month examination.

In presentation of clinical outcomes below, findings through the 24-monthexamination have been emphasized as all patients were eligible for these examinations(except for the 2 patients who died earlier in follow-up) and because theprimary outcome was defined based on the 24-month examination. Findings arepresented from other examinations to permit assessment of consistency of 24-monthoutcomes with those at other time points. Cross-sectional findings have beenpresented for each examination through 36 months because most patients (86%)were eligible for a 36-month examination. Longitudinal data displays include48-month data for patients examined at that time.

The distributions of study eyes by visual acuity at the 3-month throughthe 36-month examinations are summarized in Table4 for comparison with the baseline distributions in Table 2. Of 200 study eyes with visual acuity measured and reportedat the 24-month examination, 181 (90%) had the measurement made by a maskedtraveling vision examiner. Median visual acuity of eyes atthe 24-month examination was 20/250 in the observation arm (interquartilerange, 20/80-20/320) and 20/160 in the surgery arm (interquartile range, 20/64-20/320).The 24-month distributions of visual acuity did not differ between treatmentarms to a statistically significant degree (P = .07,Wilcoxon rank sum test). The largest difference between distributionsof visual acuity by treatment arm at any follow-up examination was observedat the 12-month examination (P = .04, Wilcoxonrank sum test) in favor of surgery.

By the 12-month examination, the visual acuity of the study eye of 1patient in each treatment arm had dropped to worse than 20/1600 (<5/400)but was at least light perception. One additional patient in the surgery armlost visual acuity to this low level by the 36-month examination; a thirdpatient in the surgery arm lost visual acuity to the level of hand motionsbased on a report from a non-SST ophthalmologist.

Distributions of changes in visual acuity from baseline to scheduledfollow-up examinations are summarized in Table5. Large improvements and large declines in visual acuity from baselinewere observed in both treatment arms at each examination. Forty-eight eyes(42%) in the observation arm and 74 (66%) in the surgery arm experienced improvementsin visual acuity of 2 lines or more at 1 examination or more after enrollment,primarily before the 12-month examination. Of these eyes, 25 (52%) of 48 inthe observation arm and 39 (53%) of 74 in the surgery arm later lost visualacuity to the baseline level or worse. Median 24-month changes in visual acuitywere a loss of 2 lines in the observation arm and a loss of 1 line in thesurgery arm (P = .14, Wilcoxon rank sumtest). The distributions of change in visual acuity differed most at the 12-monthexamination (P = .02, Wilcoxon rank sumtest, in favor of surgery). Based on 97 study eyes in the observation armand 103 eyes in the surgery arm that had visual acuity measured at the 24-monthexamination, the numbers and percentages with successful outcomes (SST definition)were 44 (45%) and 55 (53%), respectively (P = .26,χ² test). These observed rates yield a success rate of 1.18(95% CI, 0.89-1.56), that is, an estimated relative benefit of 18% for surgerybut with a confidence interval that includes 1.0, indicating the possibilityof no benefit.

Figure 1 displays the percentagesof study eyes with a visual acuity at each follow-up examination of 2 linesor more (≥ 8 letters) worse than at baseline (defined as “failures”for SST purposes), after considering both losses and recoveries of visualacuity and patients who missed the 24-month examination but had a 36- or 48-monthexamination (2-state model). The percentages of eyes with successful outcomesat the 24-month examination (and 95% CIs) calculated from the 2-state modelare 46% (36%-55%) in the observation arm and 55% (46%-64%) in the surgeryarm. These percentages yield a success ratio of 1.20 (95% CI, 0.92-1.58),nearly identical to that calculated from patients examined at 24 months. Althoughthe percentage of eyes in the surgery arm that had lost 2 lines or more ofvisual acuity (failures) was smaller than the percentage of eyes in the observationarm at each examination, the ratio of successful outcomes in surgery eyesrelative to observation eyes did not reach 1.5, the smallest success ratiodefined to be clinically significant when the trial was designed, at any examination.Furthermore, only at the 12-month examination was the 95% CI on the successratio (1.35) consistent with a benefit to surgery in the full group of patients(95% CI, 1.05-1.71).

Subgroups of patients defined by baseline characteristics that werepotential predictors of visual acuity outcomes were examined by treatmentarm for consistency with overall findings. These included age (≤45 years,>45years), gender, smoking history (ever, never), use of any hypertensivemedication, visual acuity of the study eye (≥20/100, <20/100), visualacuity of the fellow eye (≥20/20, <20/20), size of the subfoveal lesion(≤2 DA, >2 DA), earlier laser treatment (yes, no), type of CNV (classiconly, both classic and occult), and eligibility for enrollment based on strictapplication of criteria (eligible, ineligible). In 4 subgroups of patientsof the 20 subgroups considered, there was a suggestion that surgery had beenbeneficial for 2 years or longer: those older than 45 years (n = 130),women (n = 125), those using antihypertensive agents (n = 42),and patients who had visual acuity in the study eye worse than 20/100 at baseline(n = 92).

The estimated percentages of patients with a 2-line or greater lossin visual acuity for the subgroups of patients whose baseline visual acuitywas 20/50 to 20/100 and 20/125 and poorer are seen in Figure 2A and B, respectively. In the subgroup of eyes with a baselinevisual acuity of 20/100 or better (Figure 2A),median visual acuity in the observation arm dropped from 20/64 at baselineto 20/125 at the 24-month examination and in the surgery arm from 20/80 atbaseline to 20/160 after 24 months. Of 22 eyes in the surgery arm that hadbetter visual acuity at the 3-month examination (by ≥8 letters) than atbaseline, 9 had lost visual acuity to 8 letters or more worse than baselineby the 24-month examination. Median 24-month changes in visual acuity in thissubgroup were losses of 2.8 lines in the observation arm and 2.9 lines inthe surgery arm (Table 6). The successratio for this subgroup of eyes was 0.92 (95% CI, 0.61-1.40).

Among eyes in the subgroup with a baseline visual acuity worse than20/100 (Figure 2B), median visual acuityin both treatment arms dropped by 1 line, that is, from 20/200 at baselineto 20/250 at the 24-month examination. However, the interquartile range forobservation eyes was 20/200 to 20/500 while the interquartile range for surgeryeyes was 20/125 to 20/320, and the mean visual acuity was 20/320 and 20/200,respectively. Of 21 eyes in the surgery arm that had better visual acuityat the 3-month examination (by ≥8 letters) than at baseline, only 2 hadlost visual acuity to 8 letters or more worse than baseline by the 24-monthexamination. Median 24-month change in visual acuity was a loss of 1.4 lines(7 letters) in the observation arm and a gain of 0.2 line (1 letter) in thesurgery arm (Table 6; P = .08, Wilcoxon rank sum test). In this subgroup, 31 (81%)of 41 eyes in the surgery arm vs 20 (50%) of 40 eyes in the observation armhad 24-month visual acuity similar to (no more than 7 letters different from)or better than baseline visual acuity (P = .02,χ² test), yielding a success ratio of 1.53 (95% CI, 1.08-2.16).To ensure that these findings were not an artifact of the prespecified subgrouping,patients were divided into 3 subgroups based on the baseline visual acuityof the study eye. Two-year visual acuity changes suggestedno benefit and possible harm from surgery for eyes with a visual acuity of20/50 to 20/64 (success ratio, 0.85; 95% CI, 0.48-1.52), a small benefit tosurgery for eyes with a visual acuity of 20/80 to 20/125 (success ratio,1.22;95% CI, 0.69-2.14), and a large benefit to surgery for eyes with an initialvisual acuity of 20/160 to 20/800 (success ratio, 1.50; 95% CI, 1.06-2.11). In a logistic regression model with success as the dependent variableand terms for initial visual acuity, treatment arm, and interaction betweensurgery and visual acuity as independent variables, the interaction was statisticallysignificant by the conventional criterion (P = .04).Interaction also was supported when potentially predictive baseline characteristicsselected from univariate subgroup analyses (age, gender, and use of antihypertensiveagents) were included as independent variables together with treatment arm,initial visual acuity of the study eye, and a term for interaction betweensurgery and initial visual acuity of the study eye in most logistic regressionmodels (P≤.04).

The mean change in contrast threshold of study eyes from baseline tothe 24-month follow-up examination was no change in both treatment arms (Table 7). Neither the distributions of contrastthreshold (data not shown) nor the distributions of change in contrast thresholddiffered by treatment arm at any examination (P≥.15,Wilcoxon rank sum tests).

Changes in reading speed using only the study eye from baseline to follow-upexaminations (Table 8) paralleled changesin visual acuity but were more striking: changes to the 3- and 6-month examinationsfavored surgery (P<.01, Wilcoxon rank sum tests),but the benefit had diminished in comparison to observation by the 24-monthexamination (P = .07, Wilcoxon rank sumtest). In the subgroup of eyes with a visual acuity worse than 20/100 at baseline,changes in reading speed favored surgery somewhat at each examination. Bythe 24-month examination, the median change in this subgroup of eyes was aloss of 7 words per minute in the observation arm and a gain of 7 words perminute in the surgery arm (P = .10, Wilcoxonrank sum test).

Visually significant cataract, defined as either cataract surgery orlens opacity reported by the SST ophthalmologist to be sufficient to reducevisual acuity by at least 2 lines in a normal eye, developed in the initiallyphakic study eyes of 1 of 107 patients in the observation arm and 44 of 112patients in the surgery arm. All but 5 of the 44 study eyes reported to havevisually significant cataract were in the subgroup of patients 50 years orolder. No study eye in the observation arm but 27 eyes in the surgery armhad cataract surgery. Of the 27 patients who had cataract surgery in the studyeye, 26 were 50 years or older. By 24 months after enrollment,the cumulative percentage of initially phakic eyes in the surgery arm thathad had cataract surgery was 14%. Eyes in the surgery arm that hadcataract surgery had a mean change in visual acuity from baseline to the 24-monthexamination similar to the change in eyes that did not have cataract surgery:mean losses of 1.2 lines and 1.3 lines, respectively. However, following cataractsurgery, eyes in the surgery arm had smaller mean changes in visual acuityfrom baseline to the 36-month (3-letter improvement) and 48-month (3-letterloss) examinations than phakic eyes that did not have cataract surgery (remainedclose to a 1.3-line loss).

The only case of endophthalmitis developed after cataractsurgery in an eye in the surgery arm. Rhegmatogenous retinal detachments werereported in 5 eyes in the surgery arm but in no eye in the observation arm;all of the retinal detachments were repaired successfully with attachmentof the macula after a single surgery. Other ocular complications were rare. As noted earlier, 3 eyes in the surgery arm lost visual acuityto a level that could not be measured quantitatively using study methods at1 examination or more. Of these 3 eyes, 1 eye had visually significant cataractreported throughout the remaining follow-up period, beginning with the 12-monthexamination. A second patient had been followed up outside the SST since the6-month examination; when examined by the traveling vision examiner at the48-month examination, the patient was able to read 12 letters with the studyeye at a test distance of 0.5 m, corresponding to a visual acuity of 5/320(20/1280). No reason was reported for the loss of visual acuity in the thirdsurgery eye with unmeasurable visual acuity. The single eye in the observationarm that had visual acuity reported to be worse than 5/400 (<20/1600) wasfollowed outside the SST Group H Trial starting with the 12-month examination;no measurement made by an SST examiner was available after the 6-month examination.

Sixty-three eyes treated surgically had leakage of fluorescein dye fromCNV at the periphery of the postoperative area of disturbed retinal pigmentepithelium, as judged at the Photograph Reading Center, on 1 or more fluoresceinangiograms by the 24-month examination, yielding a 24-month cumulative incidenceof 58%. Of these 63 eyes, 44 eyes also had fluorescein leakage within thearea of disturbed retinal pigment epithelium. Another 14 eyes had leakageobserved only within the disturbed area. Submacular surgery was repeated in24 eyes; 18 eyes had laser photocoagulation to CNV; and 3 eyes had photodynamictherapy with verteporfin. Altogether, 42 (38%) surgery eyes had 1 or moreadditional treatments for CNV by the 24-month examination. Eyes that had fluoresceinleakage from CNV, whether located peripherally or centrally and regardlessof whether the CNV was treated, lost a mean of 1.5 lines of visual acuityby the 12-month examination and a mean of 2.5 lines by the 24-month examination.In contrast, eyes that did not have fluorescein leakage from CNV during thefirst 24 months after surgery had stable or improved visual acuity over the2-year period with a mean gain of 6 letters (1.2 lines) from baseline to the24-month examination. Six eyes in the observation arm had 1 or more treatmentsfor CNV by the 24-month examination; 2 of these 6 eyes had laser photocoagulationand 4 had photodynamic therapy. Of eyes in the surgery arm that remained freeof recurrent CNV, only 26% had lost 2 lines or more of visual acuity by the24-month examination compared with 63% of surgery eyes treated for recurrence,43% of surgery eyes that had recurrence but no subsequent treatment, and 54%of eyes in the observation arm.

Although fluorescein leakage from CNV was common after surgery, leakagefrom CNV was observed less often in eyes in the surgery arm than in the observationarm at each examination from 3 months through 24 months (Figure 3), based on reviews of photographs at the Photograph ReadingCenter. At the 24-month examination, 35% of surgery eyes (95% CI, 25%-45%)and 49% of observation eyes (95% CI, 40%-59%) had fluorescein leakage fromCNV (P = .06, χ² test).Thereafter, leakage from CNV was observed with similar frequency in eyes inboth treatment arms.

Three months after enrollment, the subfoveal lesions of more than halfthe eyes in both treatment arms were classified by personnel at the PhotographReading Center to be in a smaller size category or in the same size categoryas at baseline (Table 9). However, thesubfoveal lesions of nearly three quarters of the eyes in both treatment armshad increased in size by at least 1 category by the 24-month examination.The size of the lesion was smaller than at baseline in only 2 eyes in eacharm (P = .43, χ² test fortrend).

The SST Research Group did not observe the large benefit of surgicalremoval of subfoveal CNV compared with observation that the SST Group H Trialwas designed to detect or rule out. Promising findings reported from earlierinvestigations of submacular surgery to remove subfoveal CNV from eyes withthe ocular histoplasmosis syndrome^{23 - 25} ledto the hope that a large enough benefit would be achieved to warrant the costand perceived risks of submacular surgery. The sample size was selected toprovide sufficient power to detect or to rule out a large 2-year benefit ofsurgery, that is, 50% more eyes with stable or improved visual acuity 2 yearsafter enrollment compared with baseline visual acuity. Thus, a smaller benefitof surgery, such as 25% or 30% which many trials have been designed to detect,could be neither confirmed nor dismissed. Data provided by the SST Group HTrial suggest a better visual acuity outcome in 18% to 20% more eyes treatedwith surgery than eyes in the observation arm when all eyes enrolled are considered.This effect size is similar to that ultimately observed 2 years after randomizationof similar eyes in the SST pilot study. However, the 95% CIs on both estimatesincluded 1.0, indicating the possibility of no benefit to surgery overall.Encouraging differences between treatment arms for visual acuity and readingspeed were observed during the first year of patient follow-up (Table 4, Table 5, and Table 8 and Figure1). The high incidence of recurrent CNV and development or progressionof cataract suggest that these factors likely accounted for much of the lossof visual acuity in the second and later years in the surgery arm.

Retrospective review of data from surgical case series prior to initiationof the SST pilot study in 1993 and unpublished findings from the SST randomizedpilot trial in similar patients suggested that the subgroup of eyes with avisual acuity worse than 20/100 was the one more likely to benefit from submacularsurgery; this finding was confirmed in the SST Group H Trial. Among characteristicsof patients, eyes, and lesions at the time of enrollment, baseline visualacuity of the study eye again was found to be an important predictor of asuccessful outcome (stabilization or improvement of visual acuity) 2 yearsafter surgery in both univariate and multivariate analyses. The 2-year successratio of surgery vs observation was 1.53 (95% CI, 1.08-2.16) for study eyeswith an initial visual acuity worse than 20/100 (n = 92) and 0.92(95% CI, 0.61-1.40) for study eyes with an initial visual acuity of 20/100or better (n = 133). These CIs indicate that there was at leasta small benefit to surgery, and likely a large benefit, among eyes with aninitial best-corrected visual acuity worse than 20/100, but there was no benefitfrom surgery in eyes with an initial visual acuity of 20/100 or better. Regressionanalyses supported a qualitative interaction between surgery and initial visualacuity with respect to visual acuity outcome.

A rationale for submacular surgery was that surgery would limit thesize of the subfoveal lesion and that limiting the size of the subfoveal lesionwould lead to less loss of visual acuity. Lesions in both arms tended to enlarge,owing to continued growth of the lesions in the observation arm and recurrentCNV in the surgery arm. Although changes in lesion size, as assessed categoricallyin the SST, differed somewhat between eyes in the surgery arm and eyes inthe observation arm, with the differences favoring surgery, the differenceswere small and not statistically significant. Furthermore, initial stabilizationor improvement in the visual acuity of eyes in the surgery arm was not maintained.Findings regarding stable measurements of contrast threshold over time inboth treatment arms were consistent with the observation that the size ofthe subfoveal lesions remained at 9 DA or smaller in more than 80% of theeyes through the 48-month examination. The influence of an extrafoveal ingrowthsite, reported to portend a better visual acuity outcome,⁴⁸ wasnot evaluated by the SST Research Group.

Despite the multicenter design of the SST Group H Trial and surgicalprocedures performed by 32 different retinal surgeons with varying experiencewith submacular surgery, few serious operative complications arose. Postoperativeocular complications other than cataract also were reported infrequently.The incidence of fluorescein leakage from CNV at the periphery of the areaof disturbed retinal pigment epithelium observed following surgery on postsurgeryfluorescein angiograms by personnel at the SST Photograph Reading Center washigher than rates of persistent or recurrent CNV following laser photocoagulationreported for initially extrafoveal or juxtafoveal neovascular lesions.^{20 - 21} The posttreatment lesions followingsurgery and following laser photocoagulation were of different types and differentprotocols for management of recurrent CNV were adopted. Recurrence rates inthe SST Group H Trial were similar to those reported from a retrospectivesingle-center follow-up study of similar cases treated with submacular surgery.⁴⁹ Options for treatment of recurrent CNV were limitedto laser photocoagulation or a second surgical procedure when the SST GroupH Trial was initiated. The potential for a better outcome if photodynamictherapy with verteporfin or other treatments had been available throughoutthe study period cannot be evaluated using the SST database. If the rate ofrecurrent CNV following surgery in this patient population can be reducedby use of pharmaceutical adjuvants or other means, better visual acuity outcomeswith submacular surgery may be possible. However, a randomized trial wouldbe necessary to provide sufficient supporting evidence.

Findings from interviews regarding vision-targeted quality of life arereported in a companion article in this issue.²⁸ Theylargely mirror the visual acuity findings except that there was a trend towardimproved scores in both treatment arms overall. Self-reported visual functionfavored surgery through the 24-month interview among patients who alreadyhad bilateral CNV at baseline. However, differences between surgery and observationarms decreased thereafter for most aspects of vision-targeted quality of life.Smaller differences between treatment arms in favor of surgery were foundfor initially unilateral cases.

Treatments such as photodynamic therapy^{26 - 27} andcorticosteroids have been proposed to treat patients similar to those whoparticipated in this trial, but their effectiveness has not been documentedin randomized trials. Furthermore, the total cost of newer approaches to treatmentof CNV may not differ substantially from the cost of submacular surgery. Thus,in future clinical trials for patients eligible for the SST Group H Trialthat are designed with full knowledge of the outcomes observed by the SSTResearch Group, the treatment arm(s) to which surgery would be compared likelywould be different and a smaller difference in outcomes between treatmentarms might be specified to be clinically relevant.

Although rates of new CNV development in initially unaffected felloweyes of patients in the SST Group H Trial are yet unavailable, 89 (92%) of97 fellow eyes in the observation arm and 96 (93%) of 103 fellow eyes in thesurgery arm had a best-corrected visual acuity at the 24-month examinationthat was similar to (168 eyes) or better than (17 eyes) baseline visual acuity.Thus, 15 fellow eyes had lost 2 lines or more of visual acuity during thefirst 2 years after enrollment. Ophthalmologists should continue to monitorthe fellow eyes of patients who have a neovascular lesion in only 1 eye sothat the second eye can be treated promptly with laser photocoagulation ifCNV develops in an extrafoveal or juxtafoveal location.

Findings from the SST Group H Trial do not support submacular surgeryin similar eyes that have a best-corrected visual acuity of 20/100 or better.However, patients who have subfoveal CNV and best-corrected visual acuityworse than 20/100, as measured in the SST, and who meet other criteria forenrollment in the SST Group H trial may wish to consider submacular surgeryto improve their chances of retaining or improving visual acuity for at least2 years. Before such patients decide to undergo surgery, their ophthalmologistsshould discuss with them the findings from the SST Group H Trial and the pilottrial for similar patients that support surgery in these eyes, the small riskof retinal detachment (<5%), the larger risks of recurrent CNV (≥ 50%),the risk of cataract in older patients, the possibility that they will requireadditional treatment for these complications after submacular surgery, thestatus of the fellow eye and the long-term risk of vision loss in that eye,and the scarcity of information provided to date regarding the benefits andrisks of other treatments available.

Correspondence: Barbara S. Hawkins, PhD,SST Coordinating Center, Wilmer Clinical Trials and Biometry, 550 N Broadway,Ninth Floor, Baltimore, MD 21205-2010 (bhawkins@jhmi.edu).

Submitted for Publication: May 20, 2004; finalrevision received July 28, 2004; accepted August 3, 2004.

Writing Committee for SST Group H Trial Report No.9: Barbara S. Hawkins, PhD; Neil M. Bressler, MD; Susan B. Bressler,MD; Frederick H. Davidorf, MD; John C. Hoskins, MD; Marta J. Marsh, MS; PäiviH. Miskala, PhD; Maryann Redford, DDS, MPH; Paul Sternberg, Jr, MD; MatthewA. Thomas, MD; Cynthia A. Toth, MD. Documentation of approval of the manuscriptfor publication by all contributing individuals is on file at the SST CoordinatingCenter, Baltimore, Md.

Financial Disclosure: Individual statementsregarding financial interests are on file in the SST Chairman's Office,Baltimore, Md. Dr N. M. Bressler’s employer, The Johns Hopkins University,receives funding from Novartis Pharma AG, QLT, Inc, and Genentech, for consultingservices and research efforts by Dr Bressler. Dr S. B. Bressler’s employer,The Johns Hopkins University, receives funding from Novartis Pharma AG, andQLT, Inc, for research efforts by Dr Bressler. The terms of these arrangementsare managed by The Johns Hopkins University according to its conflict-of-interestpolicies.

Funding/Support: The SST is sponsored by theNational Eye Institute, National Institutes of Health, US Department of Healthand Human Services, Bethesda, Md, through cooperative agreements U10 EY11547,EY11557, and EY11558 with The Johns Hopkins University, Baltimore. Participatingclinical centers were supported by contracts with The Johns Hopkins University.

Acknowledgment: The Submacular Surgery TrialsResearch Group gratefully acknowledges the contributions of ophthalmologistswho referred patients to the study and, in particular, of patients who enrolled,agreed to random assignment to treatment arm, and returned for scheduled follow-upexaminations in order to benefit future patients.

Members of the Submacular Surgery Trials ResearchGroup April 1997 through September 2003 Clinical Centers and Personnel WhoContributed Data for the Group H Trial

Centers are listed in alphabetical order by city. The number of patientsenrolled is given in parentheses after the center location. Personnel listedare principal investigators and other personnel who had performed 5 or moreexaminations or procedures by the end of data collection on September 30,2003.

Clinical Centers

Emory University Eye Center, Atlanta, Ga (25):Principal Investigator: G. Baker Hubbard III, MD; PaulSternberg, Jr, MD (1997-2002); Antonio Capone, MD (1997-1999); SST Coordinator: Jayne M. Brown; Vision Examiners: Lindy G. Dubois, COMT; Judy Johnson, COMT; Natalie Schmitz; Photographers: James P. Gilman, CRA; Robert A. Myles, CRA;RaySwords.

The Wilmer Ophthalmological Institute, Baltimore,Md (22):Principal Investigator: Julia A.Haller, MD; Ophthalmologists: Peter A. Campochiaro,MD; Mark Humayun, MD; Eugene de Juan, MD; Dante J. Pieramici, MD; SST Coordinators and Vision Examiners: Michaele Hartnett, COT; PatriciaL. Hawse, COMT; Tracey L. Porter, COT; Ann Eager Youngblood, COA; Photographers: Judith E. Belt; Dennis Cain, CRA; David Emmert; RachelE. Falk; Mark Herring; Jacquelyn McDonald.

Illinois Retina Associates, Chicago and Harvey, Ill(3):Principal Investigator: Mathew W. MacCumber,MD, PhD; Ophthalmologists: Joseph Civantos, MD; KirkH. Packo, MD; SST Coordinators and Vision Examiners:Bruce L. Gaynes, OD, Pharm D; Laurie Rago, COA; Carrie Violetto; Photographers: Douglas A. Bryant, CRA; Frank Morini.

Cole Eye Institute, Cleveland, Ohio (7):Principal Investigator: Hilel Lewis, MD; Ophthalmologist: Peter K. Kaiser, MD; SST Coordinators: Laura Holody, COA; Susan C. Rath, PA-C; Larisa S. Schaaf, RN; Vision Examiner: Anthony Fattori; Photographers: Tami Fecko; Deborah J. Ross, CRA.

Retina Associates of Cleveland (2):Principal Investigator: Lawrence J. Singerman, MD; Ophthalmologist: Michael A. Novak, MD;SST Coordinatorand Vision Examiner: Kim Tilocco DuBois, COA, CCRP; Photographers: John DuBois, CRA; David Lehnhardt, COA.

Ohio State University, Columbus (17):Principal Investigator: Frederick H. Davidorf, MD; Ophthalmologist: Robert Chambers, DO; SST Coordinator: Cynthia Taylor; Vision Examiners: Jill Milliron, COA; Jerilyn Perry, COT; Photographer: Scott Savage, EMT-A.

Texas Retina Associates, Dallas (2):Principal Investigator: David G. Callanan, MD; Ophthalmologist: Gary Edd Fish, MD; SST Coordinators: Jodi R. Creighton, COA; Jeff L. Harris, COA; Nancy Resmini; RubyeRollins; Vision Examiner: Marilyn Andrews, COT; Photographers: Hank A. Aguado, CRA; Bob H. Boleman.

Duke University Eye Center, Durham, NC (20):Principal Investigator: Cynthia A. Toth, MD; Ophthalmologists: Brooks McCuen, MD; Glenn Jaffe, MD; SST Coordinators and Vision Examiners: Malcolm W. Anderson, PA-C, COT;Jennifer V. Caldwell;Photographers: Teresa JacksonHawks; Gregory C. Hoffmeyer; Jeffrey M. Napoli.

Midwest Eye Institute, Indianapolis, Ind (3):Principal Investigator: John T. Minturn, MD; SST Coordinator: Donna J. Agugliaro, RN; Vision Examiner: Shelly Cohen; Photographer: Carolyn Lamb.

University of Iowa, Iowa City (9):Principal Investigator: James C. Folk, MD; Ophthalmologist: H. Culver Boldt, MD; SST Coordinators: BettyFollmer; Steven A. Wallace; Vision Examiner: ConnieFountain, COT;Photographers: Ed Heffron, MA, CRA;Carolyn Vogel.

Mid-America Retina Consultants, Kansas City, Mo (6):Principal Investigator: William N. Rosenthal,MD; Ophthalmologist: David S. Dyer, MD; SST Coordinators and Vision Examiners: Denise Moore, RN; Barbara Petro,COT; Dalton J. Thibodeaux; Photographer: R. ScottVarner.

Southeastern Retina Associates, Knoxville, Tenn (19):Principal Investigator: John C. Hoskins, MD; Ophthalmologist: Joseph M. Googe, Jr, MD; SST Coordinators: Katie E. Carter, COA; Stephanie M. Evans; Tina ThibodeauxHigdon; Jennifer L. Holton; Vision Examiner: BruceD. Gilliland, OD; Photographers: Paul Andrew Blais;Philip Michael Jacobus.

Retina and Vitreous Associates of Kentucky, Lexington(24):Principal Investigator: William J. Wood,MD; Ophthalmologist: Rick D. Isernhagen, MD; SST Coordinators: Michelle L. Buck, COA; J. Lynn Cruz,COT; Joni D. James, RN; Jenny L. Wolfe, RN; Vision Examiners: Christine Brown, COT; Wanda Heath, COT; Catherine Millett, COA; Photographers: Marty Reid, COA; Edward Slade, CRA, COA.

Jules Stein Eye Institute, Los Angeles, Calif (1):Principal Investigator: Steven D. Schwartz,MD; Ophthalmologist: Robert Engstrom, MD.

McGee Eye Institute, Oklahoma City (1):Principal Investigator: Reagan H. Bradford, MD; Sumit K.Nanda, MD (1998-2002); SST Coordinator and Vision Examiner: Angela Monlux, COT.

Retinal Consultants of Arizona, Phoenix, Ariz (2):Principal Investigator: Jack O. Sipperley,MD; SST Coordinator and Vision Examiner: Jaclin J.Jacobsen, CRA, COA; Photographer: John J. Bucci.

Retina Vitreous Consultants, Pittsburgh, Pa (8):Principal Investigator: Robert L. Bergren, MD; SST Coordinators: Donna J. Metz, RN; Kathryn Sedory, RN; ChristinaTrombetta, CST; Vision Examiner: Linda Wilcox, COA; Photographers: David Steinberg, CRA; Alan Campbell, CRA;Gary Vagstad, CRA.

Associated Retinal Consultants, Royal Oak, Mich (1):Principal Investigator: George A. Williams,MD; SST Coordinators and Vision Examiners: KristiL. Cumming, RN, MSN; Bobbie Lewis, RN; Photographers:Patricia Streasick; Lynette Szydlowski.

Barnes Retina Institute, St Louis, Mo (32):Principal Investigator: Nancy M. Holekamp, MD; Ophthalmologists: Daniel P. Joseph, MD, PhD; Matthew A. Thomas, MD; SST Coordinators and Vision Examiners: Julie Binning, COT;Lynda Boyd, COT; Janel Gualdoni, COT; Virginia S. Nobel, COT; Photographers: Rhonda Allen; Bryan D. Barts; Jon E. Dahl; Timothy S.Holle; Deborah Kaiser, RN, COA; Ella Ort; Matt Raeber; John Mark Rogers.

West Coast Retina Medical Group, Inc, San Francisco,Calif (4):Principal Investigator: H. RichardMcDonald, MD;SST Coordinator: Margaret M. Stolarczuk,OD; Vision Examiner: Kevan E. Curren, COA; Photographers: Kelly Ann DeBoer; Sarah M. Huggans.

St Vincent Mercy Medical Center, Retina Vitreous Associates,Toledo, Ohio (17):Principal Investigator:Samuel R. Pesin, MD; Ophthalmologists: Charles K.Dabbs, MD; Nicholas J. Leonardy, MD;SST Coordinator andVision Examiner: James M. Haener, COT; Photographers: Lauren M. Cedoz, CRA; Richard D. Hill.

Resource Centers

The Wilmer Ophthalmological Institute, Baltimore:

Chairman's Office, Retinal Vascular Center:Principal Investigator and SST Chair: Neil M. Bressler,MD; Traveling Vision Examiners: Peggy R. Orr, MPH,COMT, Vision Testing Coordinator: Kristi L. Cumming,RN, MSN; Janel Gualdoni, COT; James M. Haener, COT; Michaele Hartnett, COT;Patricia L. Hawse, COMT; Economic Analyst: Eric B.Bass, MD, MPH; Other Personnel: Dawn Childs; ConnieLawson; Irene L. Felicetti (1997-2002); Patricia Staflin (1997-2001).

Coordinating Center, Wilmer Clinical Trials and Biometry:Principal Investigator: Barbara S. Hawkins,PhD; Biostatisticians: Li Ming Dong, PhD; AshleyChilds Mann, MS; Marta J. Marsh, MS; Epidemiologist:Päivi H. Miskala, PhD; Data Coordination and TelephoneInterviews: Rob G. Casper, MS; Alice D. Keith; Lee D. McCaffrey, MA(1997-2002); Dawn K. Smith (1997-2003); Systems Managementand Programming: Kurt Dreger; Harris A. Jaffee, PhD; M. Marvin Newhouse;Gregory A. Surplus; Stephen C. Grubb, MS (1997-2000); OtherPersonnel: Patricia A. James; Lisa A. Lassiter; Kelly S. Manos, MAS;Christine B. Alden (2000-2002); Takisha R. Kiah (1998-2002); Nancy A. Prusakowski,MS (1998-2000).

Photograph Reading Center, Wilmer Photograph ReadingCenter:Principal Investigator: Susan B. Bressler,MD; Ophthalmologists: Sharon D. Solomon, MD; DanteJ. Pieramici, MD (1997-2001); Srinivas R. Sadda, MD (2000-2002); Consultant: Oliver D. Schein, MD; Operations Director: Rochelle E. Smith; Judith Alexander (1997-1998); Kelly S. Manos,MAS (1998-2002); SST Coordinators: LaKaye Mbah; RevaW. Strozykowski (1997-2003); Isabel Mills (1999-2002); Photograph Graders: Rita L. Denbow, MLA; Michael P. Minotti; DeborahA. Phillips; Yan Tian.

Sponsor:

The National Eye Institute, National Institutes ofHealth, US Department of Health and Human Services, Bethesda, Md:Director: Paul A. Sieving, MD, PhD; Carl Kupfer, MD (Until2000); Deputy Director: Jack A. McLaughlin, PhD; Program Director: Maryann Redford, DDS, MPH; Mary FrancesCotch, PhD (1997-2001).

Committees and Members

Data and Safety Monitoring Committee:Voting Members (Appointed): Argye I. Hillis, PhD, Chair: Gary W. Abrams, MD; John E. Connett, PhD; ChristineGrady, RN, PhD; Earl G. Harrison, LLD (deceased); Lee M. Jampol, MD. Nonvoting Members (ex officio): Neil M. Bressler, MD; LiMing Dong, PhD; Barbara S. Hawkins, PhD; Marta J. Marsh, MS; Maryann Redford,DDS, MPH; Mary Frances Cotch, PhD (1997-2000).

Operations Committee:SSTChair: Neil M. Bressler, MD; SST Vice Chairs:Paul Sternberg, Jr, MD, and Matthew A. Thomas, MD; Members: Susan B. Bressler, MD; Barbara S. Hawkins, PhD; Maryann Redford,DDS, MPH.

Executive Committee (as of 2004):Chair: Neil M. Bressler, MD; Members: EricB. Bass, MD, MPH; Susan B. Bressler, MD; Jayne M. Brown; Hans E. Grossniklaus,MD; Julia A. Haller, MD; Barbara S. Hawkins, PhD; Nancy M. Holekamp, MD; CarolM. Mangione, MD, MSPH; H. Richard McDonald, MD; Peggy R. Orr, MPH, COMT; MaryannRedford, DDS, MPH; Paul Sternberg, Jr, MD; Matthew A. Thomas, MD; David J.Wilson, MD.

Surgery Subcommittee:Chair: Matthew A. Thomas, MD, Vice Chair: JuliaA. Haller, MD; Members: Eugene de Juan, MD; PaulSternberg, Jr, MD.

Vision Testing Subcommittee:Chair: Peggy R. Orr, MPH, COMT; Members: MichaeleHartnett, COT; Patricia L. Hawse, COMT; Marta J. Marsh, MS; Stephen C. Grubb,MS (1997-2000); Lee D. McCaffrey, MA (1997-2002); Gary S. Rubin, PhD (1997-1999).

Adverse Event Review Committee:Voting Members: Chair: Julia A. Haller, MD; Members: Gary W. Abrams, MD; Lee M. Jampol, MD; Ex OfficioMember: Barbara S. Hawkins, PhD.

Patient-Centered Outcomes Subcommittee:Chair: Carol M. Mangione, MD, MSPH; Vice Chair: Eric B. Bass, MD, MPH; Members:Neil M. Bressler, MD; Li Ming Dong, PhD; Barbara S. Hawkins, PhD; Harris A.Jaffee, PhD; Ashley Childs Mann, MS; Marta J. Marsh, MS; Päivi H. Miskala,PhD; Lee D. McCaffrey, MA (1997-2002). (Participation of Dr Mangione as chairof the Patient-Centered Outcomes Subcommittee is supported by contract betweenthe David Geffen School of Medicine, University of California, Los Angeles,and The Johns Hopkins University, Baltimore.)

Quality Assurance and Monitoring Subcommittee:Chair: Barbara S. Hawkins, PhD; ViceChair: Susan B. Bressler, MD; Members: JudithE. Belt; Li Ming Dong, PhD; Julia A. Haller, MD; Michaele Hartnett, COT; HarrisA. Jaffee, PhD; Carol M. Mangione, MD, MSPH; Marta J. Marsh, MS; PäiviH. Miskala, PhD; Peggy R. Orr, MPH, COMT; Stephen C. Grubb, MS (1997-2000);Kelly S. Manos, MAS (1998-2002); Lee D. McCaffrey, MA (1997-2002).