Landmark Studies in Uveitis

The subspecialty field of uveitis is a relatively  nascent addition to ophthalmology in general.  Worldwide, there exist only a few uveitis fellowship  training programs as few practitioners exist in general.  One of the largest training programs in the United  States has only around a hundred graduates in total  to date.¹ Being the young field it is, there exists many  debates regarding the ideal management of uveitis  patients. From early disagreements regarding physical  exam findings^2-4 to current mostly off-label use of  various steroid-sparing agents in attempts to control  disease,⁵ the field has undergone and continues to  undergo rapid growth and refinement aided by high quality research. Given the ever-increasing number  of studies being published, it may be difficult for  ophthalmologists, especially those in training, to  determine which studies are most crucial to know  concerning uveitis. As such, we have summarized  the following studies, which we believe have had  tremendous impact in uveitis and guide future  developments. It is our aim to provide the readers a  concise summary of these studies that they may gain  the valuable knowledge that these studies impart and  may hopefully, be inspired to add to the high-quality  researches discussed herein. 

Standardization of Uveitis Nomenclature for Reporting Clinical Data 

In 2004, an international workshop involving 45  uveitis specialists from 35 centers in 13 countries was  held to standardize the methods of reporting clinical  data in uveitis.⁶ The group discussed 3 aspects, namely:  (1) terminology; (2) grading of inflammation and  documentation of complications; and (3) reporting  outcomes and results. Their output was published in  the American Journal of Ophthalmology in 2005 and  has become known as the Standardization of Uveitis  Nomenclature (SUN).  

Recognizing the importance of correct anatomic  classification of uveitis to serve as the basis for  subsequent work on diagnostic criteria for various  uveitic entities, the SUN Working Group first defined  the 4 types of uveitis based on anatomic classification.  This classification scheme was adapted from the  International Uveitis Study Group, and included  anterior, intermediate, posterior and panuveitis. Next,  commonly-used terminologies to describe uveitis, in  terms of its onset (sudden vs. insidious), duration  (limited vs. persistent), and course (acute vs. recurrent  vs. chronic), were also defined.

With regard to grading of inflammation, the  group put forth separate standard methods for grading  anterior chamber cells and flare. They, however, failed  to reach consensus on a standard method for grading  vitreous cells but endorsed the grading system for  vitreous haze by The National Eye Institute with  minor modification.  

The group also enumerated the appropriate  ancillary diagnostic tests to document structural  complications in uveitis, such as fundus photography  and fluorescein angiography for optic disc and retinal  neovascularization, fluorescein angiography or optical  coherence tomography for macular edema, and so on. 

Lastly, the group defined terminologies pertaining  to uveitis activity including inactive, improved or worsened activity, and remission.

COMMENT: Although not a clinical trial, the SUN  workshop has become widely-accepted and applied  in clinical practice worldwide. It was initially intended  to standardize the nomenclature used in reporting  research study outcomes in the field of uveitis, but  has been adapted in most, if not all, ophthalmology  residency training programs as well as has been cited  in several ophthalmology and uveitis textbooks.^7,8 The  original text, which was published as a “Perspectives”  in the American Journal of Ophthalmology in 2005, is a worthwhile read with all the nitty-gritty detail often left out in textbooks. What is more exciting is that the group recently came out with 25 new publications in attempts to provide standardized classification criteria of 25 uveitic syndromes.⁹ 

Dexamethasone intravitreal implant for non infectious intermediate or posterior uveitis 

Ozurdex (Allergan, Inc, Irvine, CA, USA)  is an intravitreal, bioerodible, sustained-release  dexamethasone implant that was first approved by the United States Food and Drug Administration (US  FDA) for the treatment of macular edema associated  with retinal vein occlusion. The Dexamethasone Intravitreal Implant for Noninfectious Intermediate or Posterior Uveitis, published in 2011, was a prospective,  multicenter, single-masked, randomized, sham controlled clinical trial that determined the efficacy and safety of Ozurdex, or DEX implant, among eyes with  noninfectious intermediate and posterior uveitis over a 26-week period.¹⁰ The trial enrolled and randomized 229 patients from 46 study sites in 18 countries to 3 study groups: 77 patients received 0.7 mg DEX implant, 76 received 0.35 mg DEX implant, while the  remaining 76 patients had sham injection. Majority  (81%) of the patients enrolled had intermediate  uveitis. In all study groups, patients were allowed to  continue using their topical anti-inflammatory and/or systemic immunosuppressants under strict conditions that the doses remained stable from baseline to week 8. Primary outcome measures were the vitreous haze score and proportion of patients with vitreous haze score of 0 at week 8. Other outcome measures were time to reach a vitreous score of 0, proportion of  patients with at least two-step improvement in vitreous haze score, mean change in vitreous haze score from  baseline to week 26, best-corrected visual acuity (BCVA), and central macular thickness measured using optical coherence tomography (OCT). Safety parameters included adverse events, such as ocular hypertension, cataract, and proportion of eyes requiring rescue medications. 

Study findings showed that the proportion of  eyes with vitreous haze score of 0 at week 8 was  significantly greater in both groups that received the  DEX implant (47% in 0.7-mg DEX implant group  vs 36% in 0.35-mg DEX implant group vs 12% in  sham group). The proportion of eyes with vitreous  haze score of 0 from weeks 6 through 26 was also significantly greater in the 0.7-mg DEX implant  group than sham group. This proportion was also  significantly higher in the 0.35-mg DEX implant  group than the sham group at weeks 6 to 12 and at  weeks 20 to 26. The proportion of eyes with at least  a two-step improvement in vitreous haze score was  also significantly higher in the 2 DEX implant groups  compared to the sham group. In terms of BCVA, the mean improvement from baseline BCVA and  the proportion of eyes that achieved at least 3 lines of improvement from baseline were greater in both DEX implant groups than the sham group. OCT studies showed there were significant reductions in the mean central macular thickness at weeks 8- and 26 compared to baseline in both DEX implant groups, whereas there was no statistically significant change in the central macular thickness in the sham group. The mean change in central macular thickness was similar between the DEX implant groups. Lastly, safety analysis revealed that the proportion of eyes  requiring rescue medication at weeks 3- and 26 after  the injection was significantly higher in the sham group than both DEX implant groups. Intraocular pressure  (IOP) ≥35 mmHg was reported in less than 5% of eyes across all treatment groups and study visits, while  less than 10% of eyes had IOP ≥25 mmHg. Less than a quarter of the patients in the 0.7-mg DEX implant  group required IOP-lowering medications throughout  the 26-week study period. Of these, majority required only 1 drug to achieve IOP control. Progression of cataract was higher in the 2 DEX implant groups compared to the sham group, but this failed to reach statistical significance. The rates of other ocular adverse events were also similar among the 3 groups.  

The study concluded that a single dose of  DEX implant was effective in controlling intraocular inflammation in eyes with non-infectious forms of  posterior and intermediate uveitis. Furthermore, the 0.7-mg implant was more effective than the 0.35-mg  implant with equal safety profile.  

COMMENT: This multi-center clinical trial provided high-level of evidence on the effectiveness  and safety of Ozurdex for the treatment of non infectious intermediate and posterior uveitis. Since then, Ozurdex implant has been a valuable addition to the treatment armamentarium against non-infectious  uveitis. It is most especially beneficial for patients who  are intolerant or have contraindications to high-dose oral steroid therapy. Additional advantages include  being an easy method of administration that can be done as an in-office procedure and effectiveness lasting up to 6 months, reducing the need for repeated  localized steroid injections.  

Systemic anti-inflammatory therapy versus fluocinolone acetonide implant for intermediate,  posterior and panuveitis 

The Multicenter Uveitis Steroid Treatment  (MUST) Trial was a multicenter, randomized, controlled, clinical trial that determined whether an intravitreal fluocinolone acetonide implant (Retisert,  Bausch & Lomb, Rochester, NY, USA) was more  superior than systemic therapy of corticosteroids plus  immunosuppressants in the treatment of noninfectious intermediate, posterior and panuveitis.¹¹ From  December 2005 to December 2008, 255 patients aged 13 years and older with non-infectious intermediate, posterior or panuveitis in one or both eyes who required systemic corticosteroids to achieve uveitis  control were enrolled in 23 centers in 3 countries. One hundred twenty-nine (129) patients were randomized  in the implant group, while 121 patients were enrolled  in the systemic treatment group. Study participants in the implant group received surgical fluocinolone acetonide 0.59 mg implant in one or both eyes, followed by tapering and discontinuation of systemic corticosteroids and immunosuppressants. While the  majority of the participants in the systemic treatment  group received oral prednisone at 1 mg/kg/day or  60 mg/day. A steroid-sparing immunosuppressive was allowed, when indicated. The primary outcome was change in visual acuity (VA) from baseline to 24  months. Other outcome measures included: visual field sensitivity, clinically-graded uveitis activity, ocular and systemic complications, and quality-of-life and health utility questionnaires. 

Results of the study showed that both treatment groups had modest visual improvement of about 1- Snellen line from baseline to 24 months. There was, however, no statistically significant difference between the 2 treatment groups. In terms of uveitis control,  significantly more eyes had controlled inflammation in the implant group than in the systemic treatment  group at 24 months (88 vs 71%, respectively). In addition, eyes receiving the implant were 1.47x more likely to achieve a two-step improvement in vitreous haze at 24 months compared to the systemic treatment group.  

In terms of ocular complication, the implant group had higher rates of IOP rise, glaucoma, cataract progression, and cataract surgery compared to the systemic treatment group. Transient vitreous  hemorrhage was the most common procedure related complication. On the other hand, the risk for systemic infection requiring treatment was higher in the systemic treatment group. However, the risk of  hospitalization was similar in both groups. The rate of  hypertension was also higher in the systemic treatment  group but the rate of initiation of blood-pressure lowering medications was the same in both groups. Other systemic complications of corticosteroids  such as osteoporosis, fractures, diabetes mellitus, and  hyperlipidemia were similar in both groups. 

The authors concluded that the fluocinolone acetonide implant or systemic treatment was similarly effective in controlling intraocular inflammation. Neither was superior over the other. The choice for therapy for a patient with non-infectious uveitis should take the advantages and disadvantages of each approach into consideration. Lastly, long-term  treatment with systemic immunosuppressants was  safe and well-tolerated.

COMMENT: As of this writing, Retisert^TM is still  not available in the country. Nonetheless, two notable  findings in the study include: (1) similar effectiveness of systemic therapy with aggressive use of steroid sparing immunosuppressants and Retisert^TM; and (2) safety and tolerability of long-term treatment with  systemic immunosuppressants which is a cornerstone  for management of several inflammatory eye diseases. Furthermore, a 7-year follow-up of a cohort group of 180 patients from the original MUST study revealed  that those who received systemic therapy had better visual outcomes compared to the implant group at final visit.¹² Indications for initiation of systemic  immunosuppressants in the paper (i.e. refractory disease and certain high-risk uveitis syndromes) very much apply in the real-world clinical practice as well.

Adalimumab in Patients with Active Non-infectious Uveitis (VISUAL I) 

The VISUAL I trial, entitled “Adalimumab in  Patients with Active Non-infectious Uveitis” on its publication in September of 2016, was a multinational phase 3 trial involving 18 countries and was conducted between August of 2010 up until August of 2014.¹³ Participants included individuals aged 18 year and older with a diagnosis of active noninfectious intermediate, posterior, or panuveitis involving at least one eye. Crucially, all patients included must have had uveitis that was persistent despite the use of prednisone (10 to 60 mg per day) or an equivalent glucocorticoid for 2 or more weeks before screening. Patients with contraindications to monoclonal antibodies or those  receiving other forms of immunosuppression for systemic illness other than ocular were excluded. Patients with recent eye surgery or significant opacity precluding examination of the posterior pole were also excluded.

All patients were randomly assigned to receive either adalimumab, a fully humanized monoclonal antibody that functions as an inhibitor of tumor necrosis factor (TNF), or a placebo in a 1:1 ratio. Patients in the adalimumab group received the  standard 80-mg drug loading dose, followed by 40 mg maintenance dosing every two weeks via subcutaneous route for the duration of the study. All patients were given 60 mg prednisone at the start of their trial, which was on a preset tapering schedule that persisted until week 15, wherein all patients were placed off steroids.

The main treatment end point was either physician-determined treatment failure or persistent control of uveitis past week 80, whichever came first. The trial was originally set to run until treatment failure in 138 patients, though it eventually concluded  with 144 treatment failures as 6 additional patients were noted to have recurrence of disease by the last clinical exam. Nine ranked secondary endpoints were  recorded on each visit (i.e., change in anterior chamber cell grade in each eye, change in vitreous haze grade in each eye via fundus photo, change in BCVA in each eye via ETDRS chart, time to OCT evidence of  macular edema in at least one eye, percent change in  central retinal thickness on OCT in each eye, change in  NEI Visual Functioning Questionnaire-25 [VFQ-25] composite score, change in VFQ-25 distance vision subscore, change in VFQ-25 near vision subscore,  and change in VFQ-25 ocular pain subscore). All  patients who received adalimumab were monitored for adverse events until 70 days past their last given dose.

The trial included a total of 217 patients and found a significant median time to treatment failure  of 24 weeks in the adalimumab group versus just 13 weeks in the placebo group (Hazard ratio  [HR] 0.50; 95% CI 0.36-0.70; P<0.001) with early  and sustained separation of the treatment-failure curves. Furthermore, hierarchical testing of the ranked secondary outcomes showed that worsening  of anterior chamber cell grade, worsening of vitreous haze grade, and worsening of BCVA were significantly less common among patients who  received adalimumab (P≤0.01 for all 3 end points). VFQ-25 overall and subscore analysis also showed that the results favored adalimumab for each outcome with the exception of the change in VFQ-25 distance  vision subscore. However, the difference between  the groups in the time to OCT evidence of macular  edema was not significant. Regarding adverse events, there were no significant differences noted between adalimumab and placebo with most adverse events related to adalimumab being noted as mild, such as nausea, injection site pain, and body malaise. Though two cases of cancer (i.e., GI cancer and glioblastoma) were noted in the adalimumab group, these were  deemed by the investigator not to be secondary to adalimumab use.

The investigators concluded that adalimumab  was both safe and effective for use in non-infectious  intermediate, posterior, or panuveitis.

COMMENT: Though prior to this trial, TNF  inhibitors were already being used in the treatment  of non-infectious uveitis, particularly those that were  steroid resistant. Most evidence for this was either  based on case reports^14-16 or open label trials.^17-19 The VISUAL I trial clearly demonstrated that adalimumab  was not only superior to placebo, but also provided  a safe and prolonged period of treatment success  for these patients. This paved the way for US FDA  approval of adalimumab for the treatment of non infectious uveitis. This was a clinical milestone as  adalimumab was the first drug in its class to be  granted FDA approval for uveitis treatment.²⁰ Follow up studies to the VISUAL I trial would later further  increase indications and are discussed later on in this  review.  

Adalimumab for prevention of uveitic flare in  patients with inactive non-infectious uveitis controlled by corticosteroids (VISUAL II)

The VISUAL II trial, entitled “Adalimumab for  prevention of uveitic flare in patients with inactive  non-infectious uveitis controlled by corticosteroids  (VISUAL II): a multicentre, double-masked, randomised, placebo-controlled phase 3 trial”, was the  follow up to the VISUAL I trial published in August  2016 and involved many of the same collaborators  as its predecessor.²¹ It expanded involvement to 21  countries, 3 more than in VISUAL I. In contrast to  its predecessor, which evaluated active non-infectious  uveitis patients, VISUAL II investigated inactive,  non-infectious intermediate, posterior, or panuveitic  uveitis. Inactive uveitis was defined as clinical inactivity  for at least 28 days before the baseline visit wherein  use of oral prednisone 10–35 mg daily to maintain  inactive disease was permitted. Exclusion criteria were  identical to VISUAL I. 

All patients were randomized to either receive  adalimumab or placebo injection in a 1:1 allocation ratio. Adalimumab dosing was identical to the VISUAL I protocol. As included patients were  clinically inactive, prednisone boost was not given  at the start of trial and all patients were gradually  tapered off existing doses when applicable, with no patient being on steroids by week 16. Both primary  and secondary endpoints were identical to those in  the VISUAL I study. 

The trial included an end total of 229 patients  (114 in the placebo group and 115 in the adalimumab group) and found that there was an early and sustained  separation of the treatment failure curves between  the adalimumab and placebo groups. Treatment  failure occurred in 61 (55%) of 111 patients in the  placebo group compared with 45 (39%) of 115  patients in the adalimumab group. Furthermore, the  time to treatment failure was significantly improved  in the adalimumab group compared with the placebo  group (43% risk reduction) and more than half the  adalimumab-treated patients did not have treatment  failure vs. 8.3 months with placebo (HR 0.57, 95% CI  0.39–0.84; p=0.004). The secondary endpoint results,  however, were numerically, but non-significantly, in  favor of adalimumab for all ranked secondary variables  except for change from baseline in VFQ-25 near  vision subscore. There were no significant differences  regarding adverse events with one malignancy (non serious squamous cell carcinoma) noted and deemed  related to adalimumab use by the investigators. 

The investigators concluded that, in addition to  their previous finding supporting adalimumab use in  active uveitis disease, use of adalimumab could also  effectively allow for safe withdrawal of maintenance  steroids in clinically inactive cases without increasing  the risk of disease flare-up or increasing risk of  adverse events. 

COMMENT: The results of the VISUAL II trial  gave rationale for the eventual US FDA approval of  adalimumab to cover inactive non-infectious uveitis  as well. This landmark trial proved that not only  was adalimumab superior to placebo in regards to  controlling active disease, but that it was also a safe  and effective alternative to steroids as a maintenance  medication. This is of particular importance as long  term use of corticosteroids above a dose of 10 mg/ day carries several notable side-effects from weight  gain up to loss of bone density.²² Ophthalmologic  side-effects exist with chronic steroid use as well, such  as cataract formation and increased IOP.²³ Despite  these side effects, a subset of uveitis patients become  dependent on immunosuppressive medications, such  as steroids, in order to prevent reactivation or flare up of their disease. As such, the ophthalmologist  is often forced to balance using the lowest dose of  immunosuppressive agent needed to control the eye  disease while causing the least amount of side-effect.  Though other steroid-sparing agents exist, such as  methotrexate, mycophenolate, and azathioprine  among others, these carry their own individual side effect profiles, notably liver damage with long-term  use, and may not be suitable for every patient.²⁴ Thus, the approval of a new drug class in adalimumab, which  has excellent long-term safety profile in use by other  specialties, such as rheumatology and dermatology,^{25, 26} legitimized this option and made the use of TNF  inhibitors more accessible to the general population. 

Long-Term Safety and Efficacy of Adalimumab in Patients with Non-infectious Intermediate Uveitis, Posterior Uveitis, or Panuveitis (VISUAL III) 

Entitled “Long-Term Safety and Efficacy  of Adalimumab in Patients with Non-infectious  Intermediate Uveitis, Posterior Uveitis, or Panuveitis”  and published only last year, the VISUAL III study  was an open-label extension study of the preceding  VISUAL II and involved nearly all of the same  collaborators as the latter, comprising a total of 85  centers from 21 countries.²⁷ Any patient with non infectious uveitis qualifying for adalimumab therapy  were included. Exclusion criteria were identical to that  of the preceeding VISUAL I and II trials. 

In a natural extension to the preceding VISUAL  studies, VISUAL III investigated the ability of  adalimumab to maintain long-term quiescence,  defined as no new active inflammatory chorioretinal  vascular lesions, inflammatory retinal vascular  lesions, or both, and anterior chamber cell grade  and vitreous haze grade of 0.5. or less in both eyes  relative to baseline. As this was an open-label trial  based on clinical practice, all patients were permitted  to continue, taper, or discontinue concomitant  corticosteroid therapy, immunosuppressive therapy,  or both at the investigator’s discretion. Patients were  also allowed up to two or fewer PTA injections per  eye per year so long as quiescence was maintained. Secondary objectives were identical to the previous VISUAL trials. 

All patients were given subcutaneous adalimumab  with 80 mg as loading dose, then 40 mg every two  weeks starting one week after. Patients were then  evaluated at weeks 0-, 2-, 4-, 8-, 12-, and 18, then  every 12 weeks thereafter until the final visit. Data  were collected for up to 7 years per patient, but data  analysis was standardized at up to 150 weeks in order  to confer uniformity to the study. AE monitoring was  continued for up to 70 days after last treatment with  adalimumab.

A total of 424 patients were included in the final  analysis. At study entry, 67% of patients had active  uveitis and 33% had inactive uveitis. Quiescence was  maintained beyond week 78 in both active and inactive  groups, with 80% of patients in the active group and  96% in the inactive group showing quiescence at  week 150. Furthermore, at week 150, 54% of patients  with active uveitis at study entry and 89% of patients  with inactive uveitis achieved corticosteroid-free  quiescence. Finally, for patients with active uveitis at  study entry who were in quiescence at week 150 and  receiving systemic corticosteroids, most were receiving 7.5 mg/day or less. On subset analysis, it was noted that  out of 141 patients receiving corticosteroids to control  active uveitis at study entry, 68 remained in the study at  week 150, with 44% of those showing corticosteroid free quiescence at week 150. Regarding patients with  active uveitis at study entry, 68% experienced one or  more episodes of uveitis recurrence between week 8  and their final visit and 9% discontinued adalimumab  because of recurrence. Regarding patients with  inactive uveitis at study entry, 39% experienced  one or more episode of uveitis recurrence between  week 0 and their final visit, with 0.8% discontinuing  adalimumab because of recurrence. Overall, the  trends observed for quiescence were similar for  other efficacy variables, including AC inflammation,  vitreous haze, CST, and BCVA. The mean daily  dose of systemic corticosteroids was reduced from 9.4 – 17.1 mg/day at week 0 to 1.5 – 3.9 mg/day at  week 150 for all patients.

Adverse event (AE) profile was likewise  acceptable. While 226 patients (53% or 80 events/100  person-years [PY]) experienced one or more AEs that  were considered by the investigator to be possibly or  probably related to the study drug, no AEs resulted in  permanent blindness. The most frequently reported  AEs were infections, with 275 instances recorded (65%  or 79 events/100 PY). Thirteen (13) patients (3% or  1.3 events/100 PY) reported treatment-emergent  malignancies, but only 4 cases of nonmelanoma skin  cancers were deemed possibly related to adalimumab  use by the investigators. Six (6) patients (1.4% or  0.5 event/100 PY) reported treatment-emergent  demyelinating events, comprising demyelination  (n=2), multiple sclerosis (n=2), and optic neuritis  (n=2). Five (5) of these patients discontinued  adalimumab as a result. Four (4) patients (0.9% or  0.4 event/100 PY), all with a medical history of  sarcoidosis, reported treatment- emergent sarcoidosis.  Four (4) deaths (0.4 event/100 PY) were reported  during the entire study period, caused by B-cell  lymphoma, metastatic pancreatic carcinoma, trauma,  and brain abscess. Of these, only the brain abscess was considered by the investigator to be possibly related to  study drug.

The authors concluded that adalimumab proved  efficacious for inducing and maintaining long-term  quiescence in non-infectious uveitis patients. Further,  they noted that the AE profile was similar to that seen  in the previous VISUAL studies and ruled that this  profile was within acceptable risk levels. 

COMMENT: With the completion of VISUAL III, a  full exploration of the usage of adalimumab for non infectious uveitis was completed. Ophthalmologists  had been prior informed that this TNF inhibitor was  effective for both active uveitis in terms of controlling  disease²⁸ and for inactive uveitis in terms of enabling  steroid-tapering.²⁹ This study examined the use of  adalimumab in a real-world setting and found excellent  integration into the existing uveitis armamentarium.  As of this writing, the effects of this study have yet to  be seen as it is the most recent study to be included in  this list of uveitis landmark studies. It is hopeful that,  given the stellar overall performance of adalimumab  in this three-part investigation, the door would be  open for future approval of other drugs in this class.  Current monoclonal antibody drugs under trial for  uveitis include: baricitinib, a janus kinase inhibitor  (NCT04088409); efalizumab, a lymphocyte inactivator  (NCT00280826) and; golimumab, a TNF inhibitor  similar to adalimumab (NCT04218565). Other drugs  are also being investigated. Recently, the STOP uveitis  trial, which investigated the efficacy of tocilizumab,  an anti-IL-6 antibody medication, in the management  of active non-infectious uveitis, was completed with  favorable reports in international conventions.³⁰ With  the ever-increasing adoption and testing of new and  existing medications, the practice of uveitis continues  to evolve at a rapid rate. 

The PeriOcular vs. INTravitreal corticosteroids for uveitic macular edema (POINT) Trial

The POINT trial, entitled “Periocular Triam cinolone vs. Intravitreal Triamcinolone vs. Intravitreal  Dexamethasone Implant for the Treatment of Uveitic  Macular Edema: The PeriOcular vs. INTravitreal  corticosteroids for uveitic macular edema (POINT)  Trial”, was published in 2018 and was a randomized  multicenter trial involving three countries (USA, UK,  and Australia).³¹ The trial enrolled patients with active  or inactive, non-infectious anterior, intermediate,  posterior, or panuveitis, with a focus on all enrolled patients having macular edema (ME) on OCT  deemed to be secondary to their uveitis. Furthermore,  all patients investigated needed to have persistence  and stability of their ME findings, such that if receiving systemic medications for the treatment of  uveitis, patients needed to be on stable doses of oral  corticosteroids and immunosuppressive drugs as  applicable for at least four weeks.

The aim of the POINT trial was to compare  treatment efficacy of the different routes of steroid  administration, whether (periocular injection, intra vitreal injection, or intravitreal implant, on ME. To  that end, its primary objective was the change in  central subfield thickness (CST) on OCT at week  8 relative to the findings at baseline. Absolute  numerical changes could not be investigated as use  of different OCT machines were allowed in order to  potentially include a larger number of institutions and  subsequently, a larger number of patients. Secondary  endpoints included: change in CST at other time  points and mean change in BCVA over the entire 24  weeks follow-up. The proportion of eyes with either  improvement, defined as 20% reduction in macular  thickness or normalization of macular thickness even  if there is <20% reduction, or resolution, defined as  normalization of the macular thickness to less than 2  standard deviations above normative mean, were also  calculated over the follow-up period. Adverse events  monitored for included: need for rescue treatment,  IOP changes, and the proportion of patients  requiring glaucoma and/or cataract surgery during  the observation period. 

The investigated patients were randomized at  a 1:1:1 allocation ratio, though those with bilateral  disease were given the same treatment for both eyes  as standard ethics dictates. Periocular steroids (PTA)  were administered as 40 mg triamcinolone acetonide  given via periorbital floor or posterior sub-Tenon’s  approach, depending on the preference of the  administering physician. All intravitreal injections  (ITA) utilized 4 mg triamcinolone acetonide with the  injection site being up to the decision of the injecting  doctor. All intravitreal implants (IDI) utilized the  same 0.7 mg dexamethasone implant (Ozurdex,  Allergan, Dublin, Ireland). Patients were evaluated at  baseline and at 4, 8, 12, 20, and 24 weeks of follow up. Images were taken on all visits except on week 20.  Re-treatment was allowed at the 8-week visit for the  periocular and intravitreal triamcinolone treatment  arms and at the 12-week visit for implants.

The trial investigated a total of 192 patients (235  eyes) and found that overall, CST improved compared  with baseline at all follow-up visits for all treatment  groups (P <0.0001) with percent reductions of 23%,  39%, and 46%, for PTA, ITA, and IDI, respectively.  However, comparative analysis revealed that both  ITA (ITA/PTA, HR, 0.79; 99.87% CI, 0.65-0.96) and  IDI (IDI/PTA, HR, 0.69; 99.87% CI, 0.56-0.86) were  superior to PTA (P <0.0001) and that this superiority  was evident as early as week 4 with early separation of  the treatment curves that eventually became attenuated  by week 24. Further, IDI was superior to ITA, but  not in a clinically significant level (P=0.035). On  longitudinal comparison, it was noted that both ITA  and the IDI were superior to PTA for improvement  of the uveitic ME at all follow-up time points, except  at the 24-week visit, upon which it was deemed by  the investigators that no further medical benefits were  likely to be gained from the initial intervention. Both  intravitreal treatment groups were then compared  and were found to have had higher proportions of  eyes with resolution of uveitic ME when compared  with PTA at each follow-up visit through the 8-week  visit primary endpoint. Both intravitreal treatment  groups also had statistically significantly greater  improvements (4-7 letters better) in BCVA from  baseline relative to PTA during the initial treatment  period (4 and 8 weeks) and at the end of the follow up period, with a mean of 5 letters improvement at 24  weeks. Finally, no significant differences were found  between the intravitreal groups at any time. Safety  analysis also favored the intravitreal groups as there  were no significant differences regarding risk for IOP  rise among groups at any point in the study, though  the proportion of eyes treated with IOP medications  increased steadily throughout the follow-up period,  from 22% at randomization, to 32% at 8 weeks,  then to 39% at 24 weeks. No glaucoma or cataract  surgeries were encountered during the 24-week study  period. The investigators concluded that intravitreal  treatments, both ITA and IDI, performed superiorly  versus periocular steroids for the management of  uveitic macular edema. Though there was a recognized  effect on IOP from the intravitreal approach, this  effect was ruled as moderate and was not statistically  significant. The authors further concluded that either  intravitreal approach (ITA or IDI) were acceptable  and did not differ from each other regarding both  efficacy and safety. 

COMMENT: The POINT trial is quite interesting  as it appears to show clear superiority with intravitreal approaches for targeted steroid therapy in cases of  uveitic ME. Careful scrutiny of the data, however,  shows that while there was a clear early separation  of the treatment curves at week 4, PTA patients  maintained a slow, but steady decrease in the ME until  becoming equivalent to both intravitreal approaches  near week 24. This finding holds true regarding both  the primary and secondary endpoint measured showing  that targeted steroid therapy by any approach is still an  effective management for uveitic ME. Further, though  the trial did not note any significant AE profile with  intravitreal steroid use, it only utilized a total follow  up period of 24 weeks. Trials investigating the use of  intravitreal steroids for longer periods of time, such  as the 0.59 mg fluocinolone acetonide intravitreal  implant (Retisert, Bausch & Lomb, Quebec, Canada),  found that by 12 months, 100% of implanted phakic  patients developed visually significant cataracts and  44% developed glaucoma or ocular hypertension  requiring surgical management.³² Though these  findings did not hold true with the use of shorter acting IDI devices, prescribing ophthalmologists  must remain aware of these potential complications  should they attempt long-term therapy using IDI  devices in general. As this trial is relatively new as of  the time of writing, its effects on therapeutic trends  have yet to be seen. Though this study demonstrates  the capacity for faster and sustained recovery of uveitic  ME when utilizing intravitreal approaches, long term  safety and efficacy data regarding these 3 approaches  remains lacking and any targeted approach may still  be viable so long as the physician is aware of the  benefits and possible detriments of each approach,  deciding the approach based on available data and  informed discussions in partnership with the affected  patient. 

Effect of Corticosteroid-sparing Treatment with Mycophenolate Mofetil vs. Methotrexate on Inflammation in Patients with Uveitis (FAST)

The FAST trial, entitled “Effect of Corticos teroid-Sparing Treatment With Mycophenolate  Mofetil vs. Methotrexate on Inflammation in Patients  With Uveitis: A Randomized Clinical Trial” on its  publication in 2019, was a randomized clinical trial  involving multiple centers from 5 countries (USA,  India, Australia, Saudi Arabia, and Mexico).³³ It included patients with non-infectious intermediate,  posterior, or panuveitis affecting at least one eye  and requiring, but had not yet started corticosteroid sparing immunosuppressive therapy, who were aged  16 years or older.

The trial aim was to compare the treatment  efficacy of two commonly used steroid-sparing  immuno­modulatory agents in the management of  uveitis: methotrexate (MTX) and mycophenolate  mofetil (MMF). Its primary aim was corticosteriod sparing control of ocular inflammation at month 6,  with multiple secondary objectives including: treatment  success at 6 months by anatomical subtype of the  uveitis, treatment success at 12 months in patients  who continued their randomized antimetabolite,  number of patients who needing switching to the  other antimetabolite, BCVA at 6 months, and CST at  6 months on OCT. Treatment success was defined as  adequate control of uveitis with no flare-ups, based  on SUN definitions⁶ with allowance for minimal use  of steroids, defined as 7.5 mg or less daily equivalent  prednisone dose for oral steroids or two drops or less  daily equivalent prednisolone acetate dose for topical  steroids. The frequency and proportion of patients  experiencing AEs from their prescribed medication  were also recorded. 

Patients were randomized at a 1:1 allocation ratio,  with one group receiving 25 mg weekly oral MTX after  a 15 mg weekly 2-week trial dose to screen for drug  tolerance, and the other group receiving 1.5 g BID  MMF after a 500 mg BID 2-week trial dose also to  screen for drug tolerance. All patients were evaluated  at baseline, 2 weeks, then at every 4 weeks up to 6  months in total. Patients with treatment success at  this point then continued taking their randomized  medication for another 6 months. If treatment was  deemed a failure, patients switched to the other  antimetabolite with a subsequent 6-month follow-up. 

The trial investigated a total of 216 patients:  107 patients in the MTX group and 109 patients in  the MMF group. The main study finding was that  treatment success was achieved in 66.7% in the MTX  group vs. 57.1% in the MMF group (P = 0.20), with the  main reason for treatment failure being inefficacy of  treatment for majority of failure cases in both groups.  In regards to overall efficacy, neither drug proved  significantly superior to the other. Further, there was  no significant difference regarding change in visual  acuity between treatment groups. There was also  no significant difference regarding the CST changes  between treatment groups at six months. However,  MTX was superior regarding treatment success in  patients with posterior uveitis and panuveitis (74.4  vs 55.3%; difference, 19.1% [95%CI, 3.6%to30.6%];  OR, 2.35 [95%CI, 1.16 to4.90]; P = 0.02), but not  significantly different for patients with intermediate uveitis. For both drugs though, if maintained  successful at six months, 80% of MTX patients in  the methotrexate group and 74.1% of MMF patients  remained a treatment success at 12 months, with  the majority (50.0% for MTX and 55.0% for MMF)  discontinuing prednisone, indicating a high rate of  treatment success regardless of the chosen drug.  There was however, greater treatment success at 12  months on MTX (69.0%) in the 29 patients for whom  MMF had previously failed vs. patients in the MMF  group (35.0%) in the 20 patients in whom MTX had  failed (difference, 34.2% [95% CI, 6.6% to 52.6%];  OR, 4.2 [95%CI, 1.3 to 13.2]; P = 0.02).

The authors concluded that both drugs were  equally found to be effective as corticosteroid-sparing  agents for use in the treatment of non-infectious  uveitis. They further stated that both agents were  non-inferior to each other from a clinical standpoint.

COMMENT: Discussions regarding the superiority  of MTX versus MMF as initial drug of choice for  steroid-sparing in uveitis have been long and ongoing,  given the ready availability and relative ease of use of  both agents in multiple countries.^{34, 35} Though clinical  evidence exists supporting efficacy of either drug and  both have well-documented use in the management  of uveitis patients, comparisons have mostly been  conducted through either retrospective or open-label  studies. The FAST study shows clinical equivalence of  both drugs and, crucially, non-inferiority of either in  a randomized study. Although study findings leaned  slightly in favor of MTX on subset analysis, the  authors conceded that further studies are still required  for verification. As of this writing, the FAST trial is  still a relatively new one and its effects on prescribing  practice, if any, are still yet to be known. Doubtless to  say however, that the non-inferiority of either drug  lends credence to current prescribing patterns and  uveitis practitioners are still well within their rights to  opt for either MTX or MMF as their initial steroid sparing drug of choice following informed discussion  with their uveitis patients. 

Overall and malignancy-related mortalities among  patients with inflammatory eye disease treated  with systemic immunosuppressive therapy 

The Systemic Immunosuppressive Therapy for  Eye Disease (SITE) Study was a large retrospective  cohort study performed at 5 academic institutions in  the United States to determine whether use of specific  systemic immunosuppressant agents was associated with increased overall and cancer-related mortalities.³⁶ Medical charts of patients with non-infectious ocular  inflammatory disease examined from 1979 to 2005  were reviewed. These included patients diagnosed  with uveitis, scleritis, cicatrizing conjunctivitis of  mucous membrane pemphigoid, corneal, optic nerve,  and orbital inflammatory diseases. Persons with pre existing cancer before the start of the cohort were  excluded. Mortality incidence from 1979 to 2005 were  checked against the US national death registry using  the patients’ identifiers. A death was counted when  an exact match was found and the cause of death was obtained.  

The study included 7,957 patients seen over  68,751 visits over 14,910 person years. There were  936 deaths; of which, 230 (25%) were due to cancer. Out of 936 deaths, 323 had received systemic immunosuppressants while 613 were unexposed. Statistical analyses showed that the cohort’s overall  mortality and cancer-mortality risks were similar to the US population. With regard to the class of immunosuppressive drugs, the study results  revealed that antimetabolites, including azathioprine, methotrexate and mycophenolate mofetil, were not associated with significant increase in overall and cancer-related mortalities. Similar findings were also  observed with T-cell inhibitors (i.e. cyclosporin).  Interestingly, TNF-inhibitors, as an aggregate,  were associated with significant increases in overall mortality (HR: 1.99, 95%, CI 1.00-3.98, p=0.050)  and cancer-related mortality (HR: 3.83, 95%, CI 1.13- 13.01, p=0.031). However, estimated risk ratios for the two TNF-inhibitors, etarnecept and infliximab, were  similar in magnitude but insignificant. There was little information on adalimumab as it was only introduced  in 2005. Lastly, systemic corticosteroids and dapsone, individually, were not associated with increased risks of overall and cancer-related mortalities.

The authors concluded that patients with ocular inflammatory disease receiving azathioprine, methotrexate, cyclosporine, dapsone or systemic  corticosteroids most likely do not have increased risks  of overall and cancer-related mortalities compared  to the general population. Meanwhile, the use of TNF-inhibitors may be associated with small to moderate increased risks in overall and cancer-related mortalities. This finding should be interpreted with caution in light of the methodological limitations and confirmed in future studies. 

COMMENT: Several non-infectious ocular inflammatory diseases including uveitis require  prolonged systemic immunosuppression to prevent  flare-ups and preserve vision. Systemic corticosteroids  have limited use due to its constellation of side-effects.  Oftentimes, immunosuppressants are needed. One  major concern is the risk of increased malignancy and  malignancy-related deaths from use of these drugs.  In fact, the association of immunosuppressants  and cancer-related deaths is well-established among  organ-transplant patients.³⁷ There are, however,  differences in the subsets of population; patients  with ocular inflammatory diseases receive far lower  doses of immunosuppressants than transplant  patients. Hence, a few studies, mostly retrospective in  nature, have been performed to check for association  between immunosuppressants and cancer among  patients with ocular inflammatory disease. While, the  study by Yates et al. showed that patients receiving  immunosuppressants for inflammatory eye disease  have increased risk for malignancy, it was not powered  to allow stratification according to the classes of  immunosuppressants.³⁸ No cancer-related death was  observed in that study as well. The SITE Study above  is, by far, the largest retrospective cohort study on this  subject matter and provides additional evidence on the  safety profile of several immunosuppressive agents,  such as azathioprine, methotrexate, and cyclosporine,  in the doses used to control inflammatory eye diseases.  Results for the 3 immunosuppressants plus dapsone  and corticosteroids show that they do not pose risk  for malignancy-related deaths. However, the study  was inconclusive with respect to the TNF-inhibitors,  including adalimumab, and alkylating agents.

CONCLUSION

The above-mentioned studies have all greatly  added to the subspecialty field of uveitis. From standardizing clinical nomenclature, to supporting  the safe and effective use of medical therapy, these  studies provide a handy basic rationale for clinical  decision-making when managing the uveitis patient. 

REFERENCES 

1. What is an Ocular Immunlogist? The Ocular Immunology  and Uveitis Foundation. https://uveitis.org/patients/list-of specialists/ocular-immunologist/ (Accessed August 6, 2021).
2. Hogan MJ, Kimura SJ, Thygeson P. Signs and symptoms of  uveitis. I. Anterior uveitis. Am J Ophthalmol. 1959;47(5 Pt  2):155-170.
3. Kimura SJ, Thygeson P, Hogan MJ. Signs and symptoms of  uveitis. II. Classification of the posterior manifestations of uveitis. Am J Ophthalmol. 1959;47(5 Pt 2):171-176.
4. Nussenblatt RB, Palestine AG, Chan CC, et al. Standardization  of vitreal inflammatory activity in intermediate and posterior  uveitis. Ophthalmology. 1985;92(4):467-471.
5. You C, Sahawneh HF, Ma L, et al. A review and update on  orphan drugs for the treatment of noninfectious uveitis. Clin  Ophthalmol. 2017;11:257-265.
6. Jabs DA, Nussenblatt RB, Rosenbaum JT, Standardization of  Uveitis Nomenclature (SUN) Working Group. Standardization  of uveitis nomenclature for reporting clinical data. Results  of the First International Workshop. Am J Ophthalmol<.  2005;140(3):509-516.
7. Sen HN, Albini TA, Burkholder BM, et al. Diagnostic  Considerations in Uveitis. Sen HN ed. American Academy of  Ophthalmology 2020-2021Basic and Clinical Science COurse  Book 9: Uveitis and Ocular Inflammation, first ed. California:  American Academy of Ophthalmology, 2020, chap 5: 68-76
8. Smith WM, Faia LJ, Garg SJ, et al. Anatomic Classification of  Uveitis. Garg SJ ed. Wills Eye Hospital Color Atlas & Synopsis  of Clinical Ophthalmology Uveitis, second ed. Philadelphia:  Wolters Kluwer, 2019, chap 2: 8-14.
9. Van Gelder RN, Sen HN, Tufail A, et al. Here Comes the  SUN (Part 2): Standardization of Uveitis Nomenclature for  Disease Classification Criteria. Am J Ophthalmol. 2021;S0002- 9394(21):00279-8. DOI:https://doi.org/10.1016/ j.ajo.2021.05.006
10. Lowder C, Belfort R, Lightman S, et al. Dexamethasone  intravitreal implant for noninfectious intermediate or post erior uveitis. Arch Ophthalmol. 2011;129(5):545-553.
11. Multicenter Uveitis Steroid Treatment (MUST) Trial Research  Group, Kempen JH, Altaweel MM, et al. Randomized  comparison of systemic anti-inflammatory therapy versus  fluocinolone acetonide implant for intermediate, posterior,  and panuveitis: the multicenter uveitis steroid treatment trial.  Ophthalmology. 2011;118(10):1916-1926.
12. Writing Committee for the Multicenter Uveitis Steroid  Treatment (MUST) Trial and Follow-up Study Research  Group, Kempen JH, Altaweel MM, et al. Association Between  Long-Lasting Intravitreous Fluocinolone Acetonide Implant  vs Systemic Anti-inflammatory Therapy and Visual Acuity  at 7 Years Among Patients With Intermediate, Posterior, or  Panuveitis. JAMA. 2017;317(19):1993-2005.
13. Jaffe GJ, Dick AD, Brézin AP, et al. Adalimumab in  Patients with Active Noninfectious Uveitis. N Engl J Med. 2016;375(10):932-943.
14. Riancho-Zarrabeitia L, Calvo-Río V, Blanco R, et al.  Anti-TNF-α therapy in refractory uveitis associated with  sarcoidosis: Multicenter study of 17 patients. Semin Arthritis  Rheum. 2015;45(3):361-368.
15. Borrás-Blasco J, Casterá D-E, Cortes X, Abad FJ, et al.  Effectiveness of infliximab, adalimumab and golimumab  for non-infectious refractory uveitis in adults. Int J Clin  Pharmacol Ther. 2015;53(5):377-390.
16. Martel JN, Esterberg E, Nagpal A, et al. Infliximab  and adalimumab for uveitis. Ocul Immunol Inflamm. 2012;20(1):18-26.
17. Suhler EB, Lowder CY, Goldstein DA, et al. Adalimumab  therapy for refractory uveitis: results of a multicentre, open label, prospective trial. Br J Ophthalmol. 2013;97(4):481-486.
18. Rudwaleit M, Rødevand E, Holck P, et al. Adalimumab  effectively reduces the rate of anterior uveitis flares in patients  with active ankylosing spondylitis: results of a prospective  open-label study. Ann Rheum Dis. 2009;68(5):696-701.
19. Simonini G, Taddio A, Cattalini M, et al. Superior efficacy of  Adalimumab in treating childhood refractory chronic uveitis  when used as first biologic modifier drug: Adalimumab as  starting anti-TNF-α therapy in childhood chronic uveitis.  Pediatr Rheumatol Online J. 2013;11:16.
20. AbbVie’s HUMIRA® (adalimumab) Receives U.S. Food and  Drug Administration Approval to Treat Adults with Non Infectious Intermediate, Posterior and Panuveitis | AbbVie  News Center. Accessed August 6, 2021. https://news.abbvie. com/news/abbvies-humira-adalimumab-receives-us-food and-drug-administration-approval-to-treat-adults-with-non infectious-intermediate-posterior-and-panuveitis.htm
21. Nguyen QD, Merrill PT, Jaffe GJ, et al. Adalimumab for  prevention of uveitic flare in patients with inactive non infectious uveitis controlled by corticosteroids (VISUAL II): a  multicentre, double-masked, randomised, placebo-controlled  phase 3 trial. Lancet. 2016;388(10050):1183-1192.   
22. Strehl C, Bijlsma JWJ, de Wit M, et al. Defining conditions  where long-term glucocorticoid treatment has an acceptably  low level of harm to facilitate implementation of existing  recommendations: viewpoints from an EULAR task force.  Ann Rheum Dis. 2016;75(6):952-957. 
23. Dinning WJ. Steroids and the eye–indications and compli cations. Postgrad Med J. 1976;52(612):634-638.  
24. Antirheumatic Agents. In: LiverTox: Clinical and Research  Information on Drug-Induced Liver Injury. National  Institute of Diabetes and Digestive and Kidney Diseases;  2012. Accessed August 6, 2021. http://www.ncbi.nlm.nih.gov/books/NBK548204/ 
25. Burmester GR, Panaccione R, Gordon KB, et al. Adalimumab:  long-term safety in 23 458 patients from global clinical trials  in rheumatoid arthritis, juvenile idiopathic arthritis, ankylosing  spondylitis, psoriatic arthritis, psoriasis and Crohn’s disease.  Ann Rheum Dis. 2013;72(4):517-524. 
26. Gordon K, Papp K, Poulin Y, et al. Long-term efficacy and  safety of adalimumab in patients with moderate to severe  psoriasis treated continuously over 3 years: results from an  open-label extension study for patients from REVEAL. J Am  Acad Dermatol. 2012;66(2):241-251.  
27. Suhler EB, Adán A, Brézin AP, et al. Safety and Efficacy of  Adalimumab in Patients with Noninfectious Uveitis in an  Ongoing Open-Label Study: VISUAL III. Ophthalmology.  2018;125(7):1075-1087.
28. Jaffe GJ, Dick AD, Brézin AP, et al. Adalimumab in  Patients with Active Noninfectious Uveitis. N Engl J Med. 2016;375(10):932-943.  
29. Nguyen QD, Merrill PT, Jaffe GJ, et al. Adalimumab for  prevention of uveitic flare in patients with inactive non infectious uveitis controlled by corticosteroids (VISUAL II): a  multicentre, double-masked, randomised, placebo-controlled  phase 3 trial. Lancet. 2016;388(10050):1183-1192.    
30. Hassan M, Ormaechea MS, Sadiq MA, et al. Composite  Endpoint Outcomes of the STOP-Uveitis Study: Evaluating  the Safety, Tolerability, and Efficacy of Tocilizumab in  Patients with Noninfectious Uveitis. Invest Ophthalmol Vis  Sci. 2019;60(9):3857-3857. 
31. Thorne JE, Sugar EA, Holbrook JT, et al. Periocular  Triamcinolone vs. Intravitreal Triamcinolone vs. Intravitreal  Dexamethasone Implant for the Treatment of Uveitic Macular  Edema: The PeriOcular vs. INTravitreal corticosteroids  for uveitic macular edema (POINT) Trial. Ophthalmology.  2019;126(2):283-295.  
32. Arcinue CA, Cerón OM, Foster CS. A comparison between  the fluocinolone acetonide (Retisert) and dexamethasone  (Ozurdex) intravitreal implants in uveitis. J Ocul Pharmacol Ther. 2013;29(5):501-507.
33. Rathinam SR, Gonzales JA, Thundikandy R, et al. Effect  of Corticosteroid-Sparing Treatment With Mycophenolate  Mofetil vs Methotrexate on Inflammation in Patients With  Uveitis: A Randomized Clinical Trial. JAMA.2019;322(10):936- 945
34. Ali A, Rosenbaum JT. Use of methotrexate in patients with  uveitis. Clin Exp Rheumatol. 2010;28(5 Suppl 61):S145-150.
35. Klisovic DD. Mycophenolate mofetil use in the treatment of  noninfectious uveitis. Dev Ophthalmol. 2012;51:57-62.
36. Kempen JH, Daniel E, Dunn JP, et al. Overall and cancer  related mortality among patients with ocular inflammation  treated with immunosuppressive drugs: retrospective cohort  study. BMJ. 2009;339:b2480.  
37. Penn I. Post-transplant malignancy: the role of immuno suppression. Drug Saf. 2000;23(2):101-113. 
38. Yates WB, Vajdic CM, Na R, et al. Malignancy risk in  patients with inflammatory eye disease treated with systemic  immunosuppressive therapy: a tertiary referral cohort study.  Ophthalmology. 2015;122(2):265-273.