Study targets diabetic macular edema management

Columbia, SC-In a multicenter trial evaluating the outcomes following vitrectomy for diabetic macular edema (DME), investigators with the Diabetic Retinopathy Clinical Research Network (http://DRCR.net) found that visual acuity (VA) improvement was associated with the removal of epiretinal membrane during surgery and worse baseline VA.

The study also showed that a reduction in central subfield thickness (CST) measured by optical coherence tomography (OCT) was associated with removal of the internal limiting membrane, worse baseline VA, and thicker baseline CST.

However, the role of vitrectomy in DME management remains uncertain, said W. Lloyd Clark, MD, a private practitioner in Columbia, SC.

"This study tempers some very positive anecdotal papers touting the effects of vitrectomy surgery," Dr. Clark said, "but at the same time, it perhaps helps us identify subsets of patients who may be particularly good candidates for vitrectomy, particularly those who have undergone multiple other treatments."

Although the results do not prove definitively that vitrectomy is a safe and effective alternative to laser therapy, Dr. Clark said that the findings may be useful in the design of future studies. The prospective, observational trial enrolled 241 eyes of patients who underwent vitrectomy at 50 clinical sites. Investigators evaluated the association of preoperative and intraoperative factors with changes in VA and CST from baseline to 6 months after vitrectomy.

Univariate analysis

In the univariate analysis for VA, lens status was a significant factor (p = 0.01) only in eyes that underwent concurrent cataract surgery and vitrectomy. The median VA change was a modest gain of seven letters. Eyes that were phakic or aphakic prior to vitrectomy had a mean gain of three letters. However, eyes that were phakic at the 6-month exam had a median loss of four letters, and those that underwent lens removal after vitrectomy but prior to the 6-month exam had a median loss of two letters.

When designing the trial, http://DRCR.net investigators thought that the preoperative status of the posterior vitreous might predict the final visual outcome and expected to see particularly dramatic improvement in vision in eyes in which the vitreous remained partially attached, Dr. Clark said.

Results showed that attachment of the posterior vitreous was not associated with improvement in VA. However, it is often difficult for clinicians to determine in the clinic whether the posterior vitreous is attached or detached, and the physicians participating in this study may not have evaluated all the eyes correctly.

Baseline VA also was assessed, and as expected, eyes with the worst preoperative VA (<20/80) showed the most improvement, a median gain of four letters (p <0.001). The results for baseline OCT CST were similar.

"Patients that had a relatively thick retina prior to surgery did demonstrate improvement with surgical intervention, and patients with minimal swelling demonstrated little to no improvement," Dr. Clark said.

Eyes with a baseline OCT CST ≥400 µm had a median VA improvement of three letters (p = 0.03). Epiretinal membrane peeling resulted in a minimal improvement, a gain of three letters (p = 0.02). Eyes in which epiretinal membrane was not peeled lost a median of two letters. Although the univariate statistical model showed that several factors were statistically significant in terms of an association between preoperative findings and VA improvement, only two factors remained significant in the multivariate model: poor baseline VA and peeled epiretinal membrane.

The second part of the study examined the association between the preoperative and postoperative CST measured by OCT. Eyes with relatively thick retinas (≥400 µm) had a much higher median OCT change (–182 µm) than those with relatively thin retinas (p <0.001), Dr. Clark said.

The univariate analysis also showed that eyes with the worst baseline VA (<20/80) had a greater reduction in CST thickness. The median OCT change was –159 µm (p <0.001), compared with –90 and –58 µm in the two other baseline VA groups. In another finding, removal of the internal limiting membrane resulted in a greater median OCT change (–106 µm, p = 0.02) than was apparent in eyes in which the internal limiting membrane was not peeled (–88 µm).

The final multivariate analysis for OCT CST revealed that baseline CST (p = 0.001), baseline VA (p <0.001), and internal limiting membrane removal (p = 0.003) were all statistically significant.

FYI

W. Lloyd Clark, MD
Phone: 803/931-0077
E-mail: lclark@palmettoretina.com

Dr. Clark has no financial conflicts of interest.