Outcomes differ from trials with ‘real-world’ anti-VEGF therapy for DME

By Lynda Charters; Reviewed by Thomas A. Ciulla, MD, MBA

Anti-vascular endothelial growth factor (anti-VEGF) therapy is the first-line treatment in the United States that physicians treat patients with decreased vision due to diabetic macular edema (DME), with an overwhelming 95% choosing an anti-VEGF drug.

Dr. Ciulla

Specifically, bevacizumab (Avastin, Genentech) was the runaway favorite over ranibizumab (Lucentis, Genentech), and aflibercept (Eylea, Regeneron Pharmaceuticals), according to the 2016 Preferences & Trends Survey of the American Society of Retina Specialists.

The DRCR Protocol T indicated that among these 3 agents, patients gained an average of 12 letters of vision after 2 years. Another important finding was that patients with worse visual acuity at baseline achieved more visual acuity improvements than those with better visual acuity at baseline.

Assessing ‘real-world’ experience

However, what are the actual outcomes achieved with these drugs in DME patients outside of randomized clinical trials? This is a question Thomas Ciulla, MD, MBA, and David Williams, MD, MBA, addressed.

Dr. Williams“As we move toward a value-based health care system, we will be assessed by our outcomes,” Dr. Ciulla said.

The investigators also wanted to learn which patients have the greatest risk of visual acuity loss and how and why do these outcomes differ from randomized clinical trials, such as the DRCR protocol T?

To answer these questions, they used the Vestrum Health Retina Database, a demographically and geographically diverse data source, that contains the electronic healthcare records of hundreds of U.S. retina specialists and over 800,000 patients. The study looked at DME patients from January 2011 to April 2016, and follow-up data available up to October 2016.

All patients must have undergone a minimum of 3 monthly injections of an anti-VEGF drug. The patients were divided into 3 cohorts based on the length of the follow-up period up to 6, 12, and 24 months, Dr. Ciulla recounted. He is volunteer clinical professor of ophthalmology, Indiana University School of Medicine, and in private practice in Indianapolis.

Outcomes

Dr. Ciulla reported that among the 2,503 patients (mean age, 62 years) in the 6-month cohort, the mean number of treatments was 5.2 injections and the mean improvement was 6.1 letters of vision. Among the 3 drugs, the visual acuity improvements were similar.

In the 12-month cohort comprised of 2,315 patients (mean age, 62 years), the mean number of injections was 8.5 and the mean improvement was 6.3 letters. Again, the improvements among the drugs were similar.

In the 24-month cohort comprised of 1,041 patients (mean age, 61 years), the mean number of treatments was 14.8, and the mean improvement was 6.4 letters. Fewer than 50 patients were treated with aflibercept in this cohort. Only the bevacizumab and ranibizumab results were analyzed and found to be similar regarding the visual acuity improvements.

Patients at risk

The mean changes in visual acuity were similar among the cohorts (the patients gained about 6 letters of vision regardless of the cohort). “There were no meaningful differences among the 3 cohorts in the outcomes based on how long the patients were followed, making patient dropout from the study irrelevant,” Dr. Ciulla commented.

The results also showed that the real-world DME patients with worse baseline visual acuity had bigger visual acuity gains that those patients who had better baseline visual acuity, he added.

The most important finding was that the patients with a good baseline visual acuity of 20/40 or better fared worst in each of the 3 cohorts–losing about 2 letters of vision in each cohort.

“This finding cannot be explained by any meaningful differences in the numbers of injections within the cohorts,” Dr. Ciulla said. “The poor visual outcomes associated with better baseline visual acuity cannot be explained by a lower injection frequency.”

Hows and whys of different outcomes

The real and ideal worlds of treatment compared poorly with each other in this study, Dr. Ciulla said.

“The real-world DME treatment outcomes in the United States compared poorly to randomized clinical trials,” Dr. Ciulla explained. “At 12 months, the real-world patients gained about 6 letters of vision in contrast to the 11 letters gained by the patients in the randomized clinical trials, a difference that represents about 1 line of visual acuity.”

Regarding the real-world patients with good baseline visual acuity, they lost about 2 letters of visual acuity at 12 months compared with those who had good visual acuity in the DRCR protocol T. Those patients gained about 8 letters of visual acuity, a difference of about 2 lines of visual.

An explanation for these differences might be that real-world DME patients receive fewer injections compared to those in randomized clinical trials, about 8 compared with 10 injections, respectively, at 12 months. “This may partially explain the poor visual outcomes in the real-world patients,” Dr. Ciulla said.

In addition to that, the patients studied under the 2 different scenarios differ, in that those that are excluded from randomized trials have proliferative diabetic retinopathy or significant macular ischemia may have undergone previous laser therapy, have extremes in baseline visual acuity, and poorly controlled diabetes or comorbidities, such as severe hypertension or renal insufficiency.

“It is not surprising that the real-world patient population tends to have worse visual outcomes,” Dr. Ciulla noted.

Dr. Ciulla summarized the main points garnered from this investigation.

“The real-world experience with anti-VEGF for DME in the United States indicated that patients experience only a modest gain in visual acuity,” Dr. Ciulla explained. “The results showed an inverse relationship with the baseline visual acuity, in that those with better baseline visual acuity are at increased risk of visual acuity loss, which in part reflects a ceiling effect.

“Patient drop out from treatment did not affect the outcomes,” he added. “The outcomes are worse than those in randomized clinical trials, because of the decreased treatment intensity and worse patient characteristics.”

Thomas Ciulla, MD, MBA

e. thomasciulla@gmail.com

This article is based on a presentation that Dr. Ciulla delivered at the 2017 American Society of Retina Specialists meeting. Dr. Ciulla has no financial interest in any aspect of this report.