Journal Club

Recent Noteworthy Studies to Stimulate Discussion and Debate

■ Vitrectomy pros and cons. The March-April 2011 issue of Survey of Ophthalmology features an impressive review of the advantages and limitations of small-gauge vitrectomy by John T. Thompson, MD, of Retina Specialists in Baltimore.

Dr. Thompson covers 20-, 23- and 25-gauge vitrectomy, focusing on issues of presurgical preparation, intraoperative considerations, closing strategies, complications and advantages. For instance, because the trocars used for smaller-gauge vitrectomy are not as sharp as the MVR blades used to make 20-g sclerotomies, more force is required to insert smaller-gauge instruments. This results in IOP spikes, with some measurements reaching as high as 63.7 mm Hg. Furthermore, intraoperative retinal breaks are more common with 25-g vitrectomy, many being related to the creation of posterior vitreous detachments. At the same time, smaller gauges offer faster surgery times as well as reduced postoperative astigmatism.

Perhaps the most impressive part of Dr. Thompson's review is the literature survey that he conducted. A total of 168 journal articles published over a nine-month period were included; the sources are presented in alphabetical order to enable easier cross-referencing. The article provides a quick guide to the latest research on vitrectomy, condensed into 11 pages.

■ Treat and retreat. No one has determined the optimal dosing strategy for ranibizumab. Certainly monthly dosing will resolve CNV in most patients, but not all patients are candidates for monthly injections. A study from the United Kingdom on clinician-determined treatment with ranibizumab adds another voice to the body of opinions; it appears in the April 2011 issue of the British Journal of Ophthalmology.

Doctors in the St. Paul's Eye Unit at Royal Liverpool University enrolled 81 patients in a prospective cohort study who were given three monthly injections of ranibizumab, followed by retreatment based on criteria including best-corrected visual acuity, structural damage as seen on optimal coherence tomography, and leakage on fluorescein angiography, but ultimately at the discretion of a clinician.

Mean BCVA improved by 3.7 ETDRS letters at 12 months. While this is still a modest gain, a closer look at the data indicates that the vast majority (97.4%) of patients lost visual acuity, albeit fewer than 15 letters. The overall mean was offset by the 17.1% of patients who gained three lines or more of visual acuity.

Furthermore, when compared to the MARINA and ANCHOR phase 3 studies of ranibizumab, the mean visual acuity gains in this study are modest. The authors conclude that while three-month initiation dosing followed by as-needed treatment was effective for stabilization and maintenance of visual acuity, compared to monthly dosing as used in MARINA and ANCHOR, the dosing regimen testing here was clearly inferior in producing improvements in visual acuity.

■ Sleep studies in PDR. With patients suffering from proliferative diabetic retinopathy presenting a particular challenge for vitreous surgeons and complications of iris neovascularization and angle neovascularization having been reported as poor prognostic factors for those surgeries, preventive measures to combat these forms of neovascularization are needed. Building on their previous findings that sleep-disordered breathing was correlated with PDR, a team of ophthalmologists from Japan endeavored to determine a link between sleep-disordered breathing and these particular prognostic factors. They report their findings in the April 2011 issue of the American Journal of Ophthalmology.

The study authors conducted a cross-sectional, comparative case series of 151 patients with PDR who underwent surgery between April 2006 and April 2008. The patients were first divided into cohorts based on whether they had iris or angle neovascularization, and then all patients were subjected to sleep studies, during which pulse oximetry was conducted.

Sleep-disordered breathing was approximately 16% more common in patients with neovascularization. Furthermore, both the 4% oxygen desaturation index and the cumulative percentage of time with percutaneous oxygen saturation <90% were significantly higher in the cohort with neovascularization. In addition, patients using insulin for their diabetes were at higher risk for neovascularization of both kinds.

While the study authors cannot conclude that sleep-disordered breathing is a cause of iris and angle neovascularization in patients with PDR, they do consider it to be a possible risk factor. They urge continued study and prospective clinical trials to determine if the use of a continuous positive airway pressure device can cut this risk.

■ Chemical vitrectomy for BRVO. Plasmin and similar enzymes are under investigation as agents to enact “chemical vitrectomy” without surgery, and the use of plasmin in related disorders is also under way. A Spanish study team recently undertook a trial to determine whether this approach would be useful in branch retinal vein occlusion. Reporting in the March 2011 issue of Archives of Ophthalmology, they write that it may.

The authors undertook a prospective case series in which eight patients with diagnoses of BRVO received an intravitreal injection of 0.2 mL autologous plasma enzyme (APE) and then had BCVA and central macular thickness as measured on OCT recorded. The patients were followed for six months.

At the end of that period, mean central macular thickness had decreased from approximately 495 µm to 229 µm. BCVA similarly improved, from a mean preoperative value of 20/80 to 20/32 at six months. Both results were statistically significant.

Though the trial reported here was very small, that it was prospectively enrolled is promising. The study authors caution that larger, longer and controlled trials are needed, but for patients with BRVO who, for whatever reason, are unable to undergo vitrectomy surgery, intravitreal APE may provide a viable alternative.

■ Putting SD-OCT to the test. Spectral-domain optical coherence tomography has certainly won its share of advocates among retina specialists, but the automated segmentation procedures upon which the technology relies have not been thoroughly tested in macular disease. To address that lack, retinal physicians in Vienna and Budapest collaborated to produce a study they have published in the March 2011 issue of Retina.

Twenty-nine eyes in 29 patients with wet AMD were divided into three groups according to whether fluid had built up in the intraretinal space, subretinal space or subretinal pigment epithelium compartment. B-scans of each patient were taken with SD-OCT, using 512 x 128 x 1024 and 200 x 200 x 1024 scan patterns. Alignment errors were tabulated and graded.

The alignment error rates were 56% for the first pattern and 41% for the second. In addition, half of all scans had rendered RPE fit alignment poorly. Conversely, subretinal fluid demarcation, retinal thickness values, and RPE and internal limiting membrane boundaries were correctly measured, for the most part.

The authors of the study conclude that serious errors still arise when using SD-OCT in macular diseases, including AMD. These errors seem to emerge mainly from measurements taking place at the sub-RPE level. This finding may have particular relevance when diagnosing AMD with occult CNV.

With the increasing use of SD-OCT outcomes as endpoints in clinical trials, the authors of this study urge continued research into improving the accuracy of this imaging technology.

■ More OCT research. Elsewhere in Central Europe, scientists in Germany made a comparison among three different central volumetric protocols used by Heidelberg Engineering's Spectralis SD-OCT to determine whether they were interchangeable. Graefe's Archive of Clinical and Experimental Ophthalmology has posted the German team's findings online.

The German scientists chose one eye each from 31 healthy adults, subjecting that eye to three volumetric protocols in a single session. The three protocols were:

(1) Scanning area 6 mm x 6 mm, 25 scan pattern.
(2) Scanning area 6 mm x 6 mm, 4 scanning pattern.
(3) Scanning area 6 mm x 3 mm (vertical), 97 scanning pattern.

Mean foveal thickness as measured with the Spectralis was 224.0 ±15.8 µm and mean central retinal thickness was 280.8 ±16.5 µm. Overall, there was good protocol interchangeability. This, in part, can be attributed to the use of healthy maculas in the study. However, the authors believe that they have advanced the understanding of how retinal coverage on OCT relates to scan resolution. They suggest that, in the future, OCT examinations may be shorter in length as a result.

■ Edema before and after. Given the theory that vitreous may contribute to postsurgical macular edema in phacoemulsification, and given also the high rate of postphaco edema in diabetic patients, ophthalmologists at the Vanderbilt Eye Institute in Nashville investigated whether previous vitrectomy influenced the rate of macular edema following cataract surgery in diabetic eyes.

Publishing their findings in the British Journal of Ophthalmology's April 2011 issue, the study team conducted a retrospective review of 90 eyes in 70 patients who underwent vitrectomy. Of those 90 eyes, 37 ultimately also underwent phaco, at a mean time of 24 months following vitrectomy. Of these 37 eyes, none of them had preoperative macular edema.

A total of 13 (35%) of the 37 eyes suffered from postphaco macular edema. Slightly more than three-quarters of those eyes developed edema within six months, with the rate steadily decreasing as time went on. Most importantly, the authors found that previous clinically significant macular edema was a statistically significant predictor of postphaco edema.

Thus, they conclude, previous vitrectomy does not lessen rates of macular edema following cataract extraction. Because previous edema seems to predict future edema, prophylaxis is suggested in phacoemulsification cases. RP