Systemic safety of anti-VEGF treatment for ROP mandates further study

Reviewed by Robert L. Avery, MD

Take-home: BEAT-ROP (Bevacizumab Eliminates the Angiogenic Threat of Retinopathy of Prematurity) was too small to assess the systemic safety of intravitreal anti-VEGF injection for retinopathy of prematurity. Data from pharmacokinetic studies and retrospective clinical studies demonstrate high circulating blood levels and suggest potential effects on neurodevelopment.

Dr. AveryBEAT-ROP (Bevacizumab Eliminates the Angiogenic Threat of Retinopathy of Prematurity) demonstrated the benefit of intravitreal bevacizumab (Avastin, Genentech) to treat stage 3+ retinopathy of prematurity (ROP) in infants with posterior zone I or II disease. The issue of systemic safety remained an unanswered question, however, because of the study’s small size. 

Accumulating evidence is cause for concern and underscoring the need for more research evaluating the safety of anti-VEGF injections for ROP, according to Robert L. Avery, MD.

Dr. Avery reviewed findings from the published literature showing that the anti-VEGF agents enter into the bloodstream in physiologically relevant concentrations and have been shown to have pharmacologic effects.

“We know there are fellow eye and distant effects in babies treated with anti-VEGF injections for ROP,” said Dr. Avery, private practice, Santa Barbara, CA. “There (also) is biologic plausibility for adverse effects on lung and neurodevelopment.”

He also presented pharmacokinetics data supporting the theory that ranibizumab (Lucentis, Genentech) may be preferred over bevacizumab from the perspective of systemic safety.

“What is needed is a prospective clinical trial evaluating these therapies,” Dr. Avery added.

Pharmacokinetics of anti-VEGF

Dr. Avery noted that the investigators of BEAT-ROP underestimated the potential systemic risks of bevacizumab therapy, based on an erroneous belief that bevacizumab, as a large molecule, could not penetrate the intact retina or escape the eye except in small amounts.

Dr. Avery mentioned that he and other investigators have shown histologically that bevacizumab penetrates the full thickness retina.  He also referenced work by Kong et al who studied infants with ROP and showed that bevacizumab reached a maximum serum concentration of about 1 mcg/mL at about 2 weeks post-intravitreal injection and reduced systemic levels of VEGF for at least 60 days.

“The serum half-life of bevacizumab was about 3 weeks, similar to what we reported in adults treated for age-related macular degeneration, but the maximum serum concentration was about 1 mcg/mL, which is 9 to 10 times higher than what we measured in adults,” Dr. Avery commented.

Similarly, a study by Wu et al found that the serum VEGF concentration in infants was reduced for 2 to 3 months following intravitreal injection with bevacizumab, but not ranibizumab. Hence, bevacizumab has the more potent systemic suppressive effect of the two anti-VEGF agents.

Effect on fellow eye

After recognizing fellow eye effects when first exploring the use of bevacizumab for proliferative diabetic retinopathy (PDR), Dr. Avery reasoned that if these observed effects were real, the dose of bevacizumab being administered was much higher than that needed to produce a change in retinal neovascularization in the treated eye. Therefore, he sequentially used lower doses of bevacizumab in PDR and found that 6.25 mcg (1/200^th the usual clinical dose) could cause an effect on retinal neovascularization of PDR.  

“Injection of 6.25 mcg into the vitreous corresponds to an intravitreal concentration of about 1 mcg/mL,” Dr. Avery said. “That is also the maximum concentration that Kong et al found circulating in the blood in their study. Therefore, it is no surprise to see fellow eye effects in babies treated for ROP.

“Assuming that bevacizumab is circulating at physiologically active levels might also explain why it seems to have a longer duration of effect than ranibizumab when used to treat ROP,” he added.

Dr. Avery explained that although both anti-VEGF agents have a similar intravitreal half-life of roughly 1 week, bevacizumab persists in the bloodstream with a half-life of 3 weeks whereas ranibizumab is rapidly cleared with a 2-hour serum half-life.

“It would seem unusual for the persistently higher concentration of bevacizumab in the blood to be responsible for the longer duration of efficacy in the treated eye,” Dr. Avery pointed out. “However, in an animal model of VEGF-induced exudative retinal detachment, Miki et al found that bevacizumab had a greater effect on the fellow eye than direct injection of ranibizumab. The only way to explain that result would be if circulating bevacizumab was affecting the fellow eye.” 

In a similar fashion, he proposed that it is conceivable that circulating bevacizumab could contribute to controlling ROP neovascularization once intravitreal levels have decreased significantly.  

Systemic safety issues

There is concern about adverse pulmonary effects of intravitreal anti-VEGF injections for ROP because babies with ROP often have lung disease, and there is evidence that inhibition of VEGF inhibits lung development and surfactant production.

“In addition, Kong et al reported reductions in levels of numerous cytokines after intravitreal injection of bevacizumab in babies,” Dr. Avery said. “Those mediators may also be involved in the development of bronchopulmonary dysplasia. Furthermore, in BEAT-ROP, there were 4 deaths from lung disease in the bevacizumab arm but only one in the control group of infants treated with laser.”

The finding that blood flow in the middle cerebral artery was reduced in babies given a single intravitreal injection of bevacizumab for ROP, combined with data from studies evaluating risk of neurodevelopmental delay, raise suspicion about potential neural adverse effect.

The latter research is comprised of retrospective studies, including one with 61 babies that reported a statistically significant, 5-fold increased risk of neurodevelopmental problems in children treated with bevacizumab and laser compared to laser alone.

No statistically significant adverse effects on neurodevelopment were found when comparing babies receiving bevacizumab only versus laser alone, but a strong trend was found. However, another larger retrospective study, including 125 babies treated for ROP, found a statistically significant, 3-fold increase in severe neurodevelopmental disability in babies receiving bevacizumab compared with those treated with laser.

“But there is inherent bias in these retrospective studies where sicker babies may have been more likely to get bevacizumab and despite regression analysis to correct for imbalances, it is hard to know how much weight to put on the evidence from these trials,” said Dr. Avery.

“For people who say they have never seen systemic effects in adults treated with anti-VEGF agents, however, I offer for consideration the findings of a meta-analysis I conducted.”

Dr. Avery was referring to a published paper assessing patients with diabetes receiving monthly injections for 2 years to treat diabetic macular edema. The results showed that in this high-risk group, anti-VEGF treated patients had a 3-fold risk of death compared to the sham or laser-treated controls.

Dr. Avery concluded that anti-VEGF therapy has been helpful in ROP, but there is biological plausibility that these agents reach the bloodstream and could potentially cause systemic side effects in the developing ROP baby. Given the large difference in systemic pharmacokinetics of these agents, he proposed prospective clinical trials to compare the safety and efficacy of these agents in the treatment of ROP.

Robert L. Avery, MD

E: avery1@jhu.edu

This article was adapted from a presentation Dr. Avery delivered at the 2017 Retina World Congress, Dr. Avery is a consultant to and investigator for Genentech and a consultant to and stockholder in Novartis and Regeneron.