Targeted Retinopathy Screening Proposed in Type 1 Diabetes

Miriam E Tucker

April 19, 2017

A data-based model for individualizing retinopathy screening schedules for patients with type 1 diabetes could significantly reduce the frequency of eye exams without delaying the diagnosis of sight-threatening disease, potentially leading to savings of $1 billion in the US over the next 20 years, new research suggests.

The findings, from 30 years of fundus photography data on 1375 participants in the Diabetes Control and Complications Trial (DCCT) and its longitudinal follow-up Epidemiology of Diabetes Interventions and Complications (EDIC) study, were published in the April 20 issue of the New England Journal of Medicine by the DCCT/EDIC Research Group.

With the goal of keeping the probability of progression to proliferative diabetic retinopathy (PDR) or clinically significant macular edema (CSME) at less than 5%, David Nathan, MD, director of the Diabetes Center at Massachusetts General Hospital, Boston, and colleagues devised a screening schedule using exam intervals ranging from 4 years to 3 months, depending on the patient's current state of retinopathy and HbA1c.

"We're saying what you have determines the risk going forward," Dr Nathan told Medscape Medical News.

The findings suggest a departure from American Diabetes Association guidelines, with a new position statement issued earlier this year advising screening at no less than 2-year intervals for patients with type 1 or type 2 diabetes who have no retinopathy on initial exam and at least annually for those with any degree of retinopathy.

"We've actually demonstrated that we can change the frequency safely," Dr Nathan said.

In an accompanying editorial, Jamie B Rosenberg, MD, of the department of ophthalmology and visual sciences, Montefiore Medical Center, Bronx, New York, and Irena Tsui, MD, of the Doheny Eye Institute, University of California, Los Angeles, write that "targeted screening intervals could be a step forward in detecting diabetic retinopathy."

However, they also caution that individualizing the current screening schedule would be more complicated for patients and their doctors to follow and could therefore lead to a reduction in screening rates.

"Useful" Findings, but Insufficient to Change Guidelines?

Asked to comment, senior author of the ADA guideline Thomas W Gardner, MD, professor of ophthalmology and visual sciences at the Kellogg Eye Center, University of Michigan Medical School, Ann Arbor, told Medscape Medical News that he thinks the new findings are useful but probably insufficient to change recommendations.

"I think the data can help in talking to individual patients.…But can one make that recommendation based on a whole group of patients? I think that would be harder at this point. I think it's great that they want to save the healthcare system money, and I think the conclusions they made based on their data are valid, but the caveat is there's more to it than the data they analyzed and the system they used."

For example, Dr Gardner noted that the DCCT/EDIC cohort is a "highly motivated" group of patients who have been closely monitored.

"Patients in the real world do not always have that intensive support system, which is particularly important once they get retinopathy. If they can't get to receive the care that is needed, then that's where the system falls apart."  

And Dr Gardner pointed out — and Dr Nathan also acknowledged — that the results depend on photographic images, which aren't always used in clinical practice.

"In most places people get some photos, but usually it's just an eye exam. Even with experienced retina specialists with dilated pupils and cooperative patients, we can miss things. The photographic approach is clearly superior."  

Reduction in Number of Screenings and Cost: Savings of $1 Billion?

In their paper, Dr Nathan and colleagues explain that during the DCCT, standardized stereoscopic seven-field fundus photographs were taken every 6 months and the same was done every 4 years during the follow-up EDIC.

Retinopathy states were classified as state 1 (no retinopathy), state 2 (mild nonproliferative diabetic retinopathy, including microaneurysms only), state 3 (moderate nonproliferative diabetic retinopathy), state 4 (severe nonproliferative diabetic retinopathy), and state 5 (PDR, CSME, or previous self-reported treatment with panretinal or focal photocoagulation, intraocular glucocorticoids, or anti–vascular endothelial growth factor [VEGF] intravitreal injections).

The study goal was to determine a retinopathy screening schedule that would limit to less than 5% the risk of progression from states 1 through 4 to state 5 during the time interval until the next retinal exam, when progression could be detected and treatment given.

On a practical level, Dr Nathan and colleagues determined that the results suggest screening intervals of 4 years, 3 years, 6 months, and 3 months for initial retinopathy states of 1 through 4, respectively.

Higher mean HbA1c levels were also associated with a significantly increased risk for worsening retinopathy.

For example, the risk of progression from no retinopathy to state 5 was 1.0% over 5 years among patients with an HbA1c of 6% as compared with 4.3% over 3 years among those with an HbA1c of 10%.

Thus, Dr Nathan and colleagues calculate that for patients with HbA1c of 6%, this translates into scheduled examinations after 5 years for those with states 1 and 2 retinopathy, after 6 months for those with state 3, and after 3 months for those with state 4.

And for those with a mean HbA1c of 10%, scheduled screenings should be at 3 years, 2 years, 3 months, and 1 month for states 1 through 4, respectively.

They provide a user-friendly Web application that can calculate the required time interval to next retinal exam based on the two factors of current retinopathy status and mean HbA1c levels

Over 20 years, the new model results in an average decrease of 10.7 retinal exam visits per patient, a 58% drop overall compared with routine annual eye exams, say Dr Nathan and colleagues.

Given the approximate $200 cost of digital photography, roughly 1 million people with type 1 diabetes in the United States, and the current distribution of retinopathy states, the model would reduce the 20-year costs by about $1 billion, a 43.4% reduction compared with routine annual screening, they predict.

What Does the Model Leave Out?

In response to the editorialists' point about increasing scheduling complexity, Dr Nathan countered that automated computerized messaging is routine now and could overcome that problem, noting: "We seem to remember to do colonoscopy every 10 years."

In their editorial, Dr Rosenberg and Dr Tsui also noted that the indications for treatment of diabetic retinopathy may be expanding beyond just PDR and CSME because of the expansion of use of anti-VEGF injections. Dr Nathan said that the model allows for that. "If doctors say they want to treat at earlier stages, we can easily reanalyze the data."

Dr Nathan also argued that, in fact, many ophthalmologists are already adjusting the frequency of visits for individual patients based on their specific clinical characteristics. "This is what ophthalmologists do anyway….This actually provides the data."

But Dr Gardner cautioned that while the authors accounted for factors such as age, sex, diabetes duration, smoking status, hypertension, and hyperlipidemia, there are other risk factors they didn't include, such as renal failure and foot ulcers. Moreover, he said, there is also value in providing patients peace of mind.

Dr Gardner said that while he sometimes advises a low-risk patient to wait 2 years before returning, "I don't tell anybody to come back in 3 years. That's where it's not just a retinopathy issue. As physicians, we take care of patients and not just their organs."

Dr Nathan has no relevant financial relationships. Disclosures for the coauthors are listed in the paper. Dr Rosenberg reports grant support from Research to Prevent Blindness. Dr Tsui has no disclosures. Dr Gardner is a consultant for Novo Nordisk and KalVista.

For more diabetes and endocrinology news, follow us on Twitter and on Facebook.

N Engl J Med. 2017:376;1507-1516, 1587-1588. Article, Editorial


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.