Evidence for Step Therapy in Diabetic Macular Edema

Step therapy is a treatment approach often required by health insurers to control costs. It involves treating first with a lower-cost medication and, only if that proves ineffective, switching to another, more costly medication. Although step therapy is commonly applied in treating hypertension, diabetes, cancer, and other conditions, concerns regarding its potential for long-term harm to the patient have led some to view it as “fail first” therapy.^1,2 One valid criticism of step therapy is a lack of an evidence base for its efficacy and safety. In this issue of the Journal, Jhaveri, Glassman, and colleagues³ describe a rigorous trial of step therapy to treat a prevalent, potentially blinding condition.

This multicenter, randomized clinical trial was conducted by the Diabetic Retinopathy Clinical Research Retina Network, which includes more than 160 academic and community-based clinical sites and more than 500 ophthalmologists who treat retinal disorders throughout the United States and Canada. This network previously reported the results of a “non-step” trial showing that aflibercept resulted in greater mean improvement in visual acuity than bevacizumab in the treatment of diabetic macular edema when the initial loss in visual acuity was 20/50 or worse.^4,5

The design, methods, and conduct of the current trial are beyond reproach. The authors randomly assigned 312 eyes (in 270 participants) with diabetic macular edema and a visual acuity of 20/50 or worse to either first be treated with the relatively inexpensive bevacizumab (cost, $50 per dose) or first be treated with aflibercept ($1,941 per dose),⁶ with a protocol for switching from bevacizumab to aflibercept therapy if explicit criteria were met. The findings show similar outcomes in the two groups with regard to improvement in visual acuity and a thickness measurement of the central retina over the 2-year trial period.

Although clinical trials provide a high level of evidence, several factors require consideration when these findings are applied to clinical practice. The enrollment criteria in the current trial reflect typical patients with diabetic macular edema being treated in the community. Participants in each treatment group received six monthly injections initially (unless criteria for success and sustained stability were met), and then an automated algorithm determined whether additional injections were warranted. This approach differs somewhat from the administration schedule of aflibercept recommended by its manufacturer, which includes injections every 4 weeks for the first five injections, followed by injections every 8 weeks. Follow-up visits took place every 4 weeks during the first year and then every 4 to 16 weeks during the second year, depending on the examination findings. This frequency of the trial participants’ follow-up visits may be greater than that of patients’ visits in usual practice. The software that was used to determine whether reinjection was warranted is not currently available to treating ophthalmologists, and the authors appropriately note that their findings are dependent on the switching criteria they used. However, the criteria for switching from bevacizumab to aflibercept therapy were based on measurements of retinal central subfield thickness and visual acuity, which are commonly obtained in clinical practice. Overall, this trial presents evidence that applies well to clinical practice.

Switches from bevacizumab to aflibercept treatment during follow-up were frequent. At 24, 52, and 104 weeks, switching criteria were met in 39%, 60%, and 70% of the eyes, respectively, in the bevacizumab-first group. Even so, the cost savings to the health care system would probably be substantial if this treatment approach is widely embraced. A critical look at the effect on health care expenditures on the basis of the results of this trial will be helpful.

Perhaps the most important concern with the use of the bevacizumab-first strategy is the fact that the frequent follow-up of the participants in the first year of this trial exceeds what usually takes place in clinical practice. This concern is especially pertinent for persons with diabetes, who often have numerous coexisting conditions that make it more difficult for them to adhere to frequent follow-up visits. Missed or delayed follow-up visits in clinical practice could result in failure to identify vision loss at or near its et and to implement a prompt switch from bevacizumab to aflibercept therapy. Whereas aflibercept was not cost-effective relative to bevacizumab for the treatment of diabetic macular edema in a previous evaluation by this research network,⁷ it is conceivable that some patients who receive bevacizumab first will have irreversible vision loss that might have been prevented with a prompt switch to aflibercept therapy. The availability of large databases such as the Intelligent Research in Sight clinical registry of the American Academy of Ophthalmology⁸ will enable researchers to produce supportive data on the benefits and risks of bevacizumab-first treatment for diabetic macular edema, if bevacizumab-flrst treatment is embraced by ophthalmologists who treat patients with newly diagnosed diabetic macular edema.