CONTROVERSIES IN CARE

Systemic Safety of Intravitreal Anti-VEGF Therapy: Are Concerns Real or Theoretical?

EDITED BY MICHAEL COLUCCIELLO, MD

The treatment of common retinal conditions, such as neovascular age-related macular degeneration, diabetic retinopathy and retinal venous occlusive disease, has been transformed by the introduction of repeated intravitreal doses of agents that target VEGF signaling.

Concern about the systemic effects of intravitreal anti-VEGF exists in the vulnerable population that retina specialists treat: hypertension, thromboembolic events, and gastrointestinal perforation^1-6 have been reported with intravenous administration of VEGF inhibitors.

The commonly used intravitreal anti-VEGF agents do transit rapidly into the bloodstream after administration⁷ — but does this transit translate into risk for adverse systemic effects?⁸

Intravitreal dosing of bevacizumab (Avastin, Genentech, South San Francisco, CA) and aflibercept (Eylea, Regeneron, Tarrytown, NY), in contrast to ranibizumab (Lucentis, Genentech), exhibits fairly long systemic exposure and produces a marked reduction in plasma free VEGF.^7,9

Bevacizumab may be taken up into and activate platelets, and in contrast to the other anti-VEGF molecules described, it may form high molecular weight immune complexes, which can aggregate in tissues such as the renal glomeruli.^10,11 Therefore, bevacizumab may theoretically be actively involved in thromboembolic events.

Although intravitreal ranibizumab is quickly cleared from the systemic circulation,⁷ the MARINA trial suggested that intravitreal use of ranibizumab was associated with a small increase in nonocular hemorrhage risk.¹²

There has been some evidence regarding the safe usage of intravitreal anti-VEGF agents, fortunately. In the CATT study comparing bevacizumab and ranibizumab, no differences in stroke incidence, myocardial infarction incidence, or induced hypertension risk were detected (although the two-year CATT data showed that the proportion of patients with one or more serious systemic adverse events was significantly greater with bevacizumab than with ranibizumab [39.9% vs 31.7%; P=.004], and there were significantly more gastrointestinal disorders in patients treated with bevacizumab compared to ranibizumab).^12,13

No significant increases in the risks of systemic cardiovascular and hemorrhagic events or in overall mortality, cardiovascular mortality, or stroke were found in a meta-analysis involving 9,557 patients over 21 trials involving the administration of intravitreal anti-VEGF agents.¹⁴

In addition, no significant differences among the DRCRnet’s Protocol T study groups using bevacizumab, ranibizumab, and aflibercept were noted regarding the rates of serious adverse events (P=.40), hospitalization (P=.51), death (P=.72), or major cardiovascular events (P=.56).¹⁵

Unfortunately, treatment trials designed to assess the prevention of vision loss caused by these retinal conditions are inadequately powered for detecting rare, serious, systemic side effects in patient populations that possess multiple, often interrelated, medical comorbidities. Therefore, we continue to maintain surveillance as we use these agents more frequently.

This month, we are fortunate to have the perspective of Robert Avery, MD, of California Retina Consultants and the Neuroscience Research Institute of the University of California-Santa Barbara.

REFERENCES

Are There Systemic Safety Concerns With Anti-VEGF Injections?

ROBERT L. AVERY, MD

Anti-VEGF therapy has revolutionized the treatment of retinal disease. The small doses used for eye disease seem to be safe, but these agents are very potent, and there are several lines of evidence that imply that these small doses could potentially have a systemic effect.¹

As an early adopter of these injections, I was particularly concerned about possible side effects. Soon after starting to use bevacizumab for proliferative diabetic retinopathy, we noted and published fellow eye effects.²

Subsequently, I have seen fellow eye effects with ranibizumab and aflibercept as well. Numerous cases have now been reported, from reduction in diabetic or uveitic macular edema, to decreased neovascularization in retinopathy of prematurity, to a case of diabetic traction retinal detachment crunch in a fellow eye after injection.^3-7

For these effects to be due to the injection, it would seem likely that the drugs would need to pass into the bloodstream at concentrations high enough to significantly reduce VEGF levels.

There are now numerous studies showing reduced systemic VEGF levels after intravitreal injections, and we have also shown elevated levels of the anti-VEGF drugs in the bloodstream, which correlate with the decreased VEGF levels.^8-12

Even if the systemic VEGF levels are reduced after intravitreal injections, many would wonder what difference that could make with regard to safety — especially because systemically administered anti-VEGF agents at much higher doses are tolerated by cancer patients.

Indeed, individual registration trials of the intravitreal agents have not shown a significant increased risk of arteriothrombotic events (ATEs) or death; however, these trials were not powered to detect a difference in uncommon events.

However, the CATT trial found an increased risk of systemic serious adverse events (SAEs) in the bevacizumab-treated patients vs the ranibizumab-treated group.¹³ Although a meta-analysis published last year of nine AMD trials did not confirm this increased incidence in SAEs,¹⁴ I have performed a more current update of this analysis and confirmed the original finding of a significant risk with bevacizumab (manuscript under peer review).

This finding is consistent with the pharmacokinetic data that show much higher systemic exposure from bevacizumab (and aflibercept) than ranibizumab, as well as a much more pronounced reduction in plasma and serum VEGF.^9,12

What Does It Mean?

Assuming that these agents could reduce circulating VEGF levels enough to create side effects, where might we see them? Fortunately, we do not see them very often, given the low incidence of serious events seen in numerous prospective trials.

Clues to the potential risks can be obtained from systemic experience. For instance, we know that systemic administration can cause an increased risk of ATEs and wound healing complications.¹⁵ One meta-analysis of AMD trials using ranibizumab showed an increased risk of cerebrovascular accidents (CVAs) only if patients were stratified regarding their baseline risk of CVA.¹⁶

A more recent, larger meta-analysis of AMD patients demonstrated an increased risk of CVA but not of death with more intensive ranibizumab treatment, for instance, with higher doses or more frequent (monthly) treatment.¹⁷

Given the higher baseline risk for ATEs in patients with diabetes, most DME trials have excluded patients with recent CVAs or myocardial infarctions. Nevertheless, a broad meta-analysis of more than 6,500 patients in trials of ranibizumab for AMD, RVO, and DME found imbalances in wound healing, CVA, and death, but only in the DME patients.¹⁸

Several recent meta-analyses of patients treated for DME have not shown a statistically significant increased risk of ATEs or death across all treatment regimens, but one did raise concerns about a dose-dependent increased risk of death in a subset of patients.^19-21 I have performed a focused meta-analysis of patients undergoing intensive anti-VEGF therapy for DME, and found a similar risk (manuscript under peer review).

Three Points to Consider

In summary, it seems that intravitreal anti-VEGF agents are well tolerated in the vast majority of patients; however, there are several lines of evidence that give me concern for their use in at-risk patients.

First, fellow eye effects have been observed with all three major agents and across many disease states. Second, the pharmacokinetic data show elevated systemic drug levels that are consistent with the observed decrease in circulating VEGF, as well as the increased risk of SAEs in the updated meta-analysis of AMD trials. Third, several (but not all) meta-analyses have pointed to potential safety signals in at-risk populations in AMD and DME.

Hence, I believe it is prudent to consider the patient’s medical history and to assess the risk before injecting. Remember, most DME trials excluded patients with a recent ATE, so safety in this population was not evaluated in the majority of our DME trials. Therefore, it is difficult to be certain of the safety when treating this population. RP

REFERENCES