Pharmacological approaches
Currently, there are two dominant modalities for the treatment of ROP, (1) intravitreal injection of anti-VEGF antibodies and/or (2) laser ablation of the avascular retina. Laser therapy requires prolonged sedation and, even when initiated early as in the Early Treatment of ROP (ETROP) study, cannot prevent retinal detachment in 12% of cases [
45]. Thus, increasing focus has been given to a pharmacological approach which would have the benefit of bedside administration, reduced cost and would reduce the burden on specialist ophthalmologists experienced in delivering laser treatments [
46].
The antibody bevacizumab binds to all forms of VEGF-A [
47]. Intravitreal injection of bevacizumab has been widely used in other vasoproliferative ocular diseases such as age-related macular degeneration, diabetic retinopathy, central and branch retinal vein occlusion in adult patients [
48‐
51]. In addition, intravitreal bevacizumab has been used in children for conditions such as Coats disease and subretinal neovascular membranes. However, its use in all ocular diseases in both adults and children is off-label. Study data shows that, in the UK, Bevacizumab is used by some ophthalmologists to treat A-ROP and posterior type-1 ROP [
4].
BEAT-ROP, a multicentre RCT, compared the efficacy of bevacizumab against standard laser treatment, using recurrence of ROP before 54 weeks postmenstrual age as a primary outcome [
52]. The authors concluded that the anti-VEGF agent was shown to be effective against stage 3 + disease in zone I, but not in zone II and larger numbers of participants were required to comment on safety. However, these conclusions have been questioned, since using recurrence at 54 weeks as a primary outcome may miss later reactivation of ROP giving a false impression of bevacizumab’s efficacy [
53]. The figure of 54 weeks was based on the recurrence of ROP following laser treatment for ROP from the ETROP study but does not take into account the possibility of late reactivation of ROP that is seen in babies treated with anti-VEGF (see below).
A Cochrane Review [
54] of anti-VEGF agents for the treatment of ROP included four RCTs (including BEAT-ROP) comparing bevacizumab with laser therapy. The authors concluded that aggregated data suggest bevacizumab as monotherapy does not reduce recurrence or risk of retinal detachment (although, as suggested by BEAT-ROP, treatment for zone I ROP is promising).
Safety concerns have been raised by clinicians regarding the use of anti-VEGF in this age group [
53,
55]. Studies in humans show that bevacizumab, when injected intra-ocularly, can escape into the general circulation [
56]. Furthermore, primate models demonstrate that the serum concentration of bevacizumab remains raised for up to 8 weeks [
57]. The 2018 Cochrane review [
54] of anti-VEFG agents concluded that insufficient safety data existed for the recommendation for routine use of bevacizumab. Roohipoor et al. [
58] recruited 116 participants, randomised to 0.625 mg bevacizumab or laser therapy. The study found that ROP regressed in all cases of the bevacizumab arm (5 eyes—3.2%—requiring retreatment at a mean of 6.01 weeks after initial treatment) and 97.3% of eyes in the laser therapy arm (
p = 0.20). Despite follow-up to 90 weeks, only ocular complications were recorded in the bevacizumab arm (one instance of cataract formation).
The dose of bevacizumab used in Roohipoor’s and the BEAT-ROP study was 0.625 mg, half of the adult dose. However, this may be significantly more than required to neutralise VEGF in the vitreous. Further studies looking at lower doses of Bevacizumab ranging from 0.16 to 0.004 mg have shown promising outcomes [
59,
60], although there is a likelihood of reactivation occurring sooner with lower doses [
61].
An alternative anti-VEGF antibody, Ranibizumab, a human monoclonal antibody fragment derived from the parent antibody Bevacizumab [
47] has been licenced and approved for treatment of retinal vasoproliferative disorders in adults and also for ROP by the National Health Service (NHS England) [
62] and European Medicines Agency (EMA) [
63] following completion of the RAINBOW trial [
64]—225 participants, randomized to ranibizumab 0.2 mg, ranibizumab 0.1 mg, and laser photocoagulation therapy. Analysis showed that ranibizumab 0.2 mg may be superior to laser therapy (treatment success odds ratio compared to laser therapy = 2.19. 95% CI 0.99–4.82). Stahl et al. record that “Death, serious and non-serious systemic adverse events, and ocular adverse events were evenly distributed between the three groups” at 24 weeks [
64]. One notable advantage that Ranibizumab may have over Bevacizumab is the reduced duration of systemic VEGF suppression [
56,
65,
66].
Longer-term follow-up from the CARE-ROP study [
67] which compared different doses of ranibizumab (0.2 mg and 0.12 mg) shows no significant difference in ophthalmic and neurodevelopmental outcomes between the treatment arms after 1 and 2 years, respectively.
A third agent, Aflibercept, has the potential to prevent VEGF-driven pathogenesis by binding VEGF-A, VEGF-B as a soluble decoy receptor [
68]. Aflibercept is licensed for use in adult retinal vasoproliferative disorders by the US Food and Drug Administration FDA [
69] and EMA [
70]. The FIREFLEYE study tested Aflibercept in a multicentre, international non-inferiority RCT for treatment of type 1 ROP and A-ROP with an active control comparator group of babies treated with laser therapy [
71]. As with the RAINBOW study, the primary outcome was the proportion of infants without active ROP and unfavorable structural outcomes (retinal detachment, macular dragging, macular fold or retrolental opacity) 24 weeks after starting treatment. One hundred eighteen infants were randomized, and 113 were treated in a 2:1 ratio of Aflbercept (0.4 mg) (75 babies) or laser (38 babies). Although treatment success was slightly higher in the Aflibercept-treated arm, the credible interval for the treatment difference was not greater than the prespecified value of − 5% and therefore non-inferiority could not be proven. Further and larger studies of Aflibercept are therefore required to reach any definitive conclusions regarding the comparative effect of Aflibercept to laser in type 1 ROP.
ROP regression following anti-VEGF is a notable for its rapidity. The RAINBOW study documented plus disease regressing significantly faster, at a median time of 4 days (interquartile range (IQR) of 3–8 days) following 0.2 mg of Ranibizumab compared to a median time of 16 days (IQR of 9–22 days) following laser. Similarly, stage 3 was noted to regress at a median of 8 days (IQR 5–17 days) following 0.2 mg of Ranibizumab versus a median of 16 days (IQR 9–29 days) following laser. Indeed, A-ROP was noted to take nearly 3 times longer to regress following laser treatment versus Ranibizumab [
72]. This study also showed that incomplete regression of ROP requiring retreatment was more frequently observed in laser-treated eyes with around 22% of eyes requiring further treatment at a median interval of 15 days post-laser compared to 8% of Ranibizumab-treated eyes at a median interval of 21 days after injection [
72].
Significant concerns around treatment with anti-VEGF include the prolonged follow-up required to identify re-activation of the disease, that is, a recurrence of acute phase features of ROP, and the management of any persistent avascular retina in the periphery, which arises from interruption of vascular development following intravitreal anti-VEGF administration. Reactivation may occur after complete or incomplete regression of ROP after anti-VEGF injections. It rarely occurs after complete regression of ROP following laser treatment. Reactivation following anti-VEGF treatment may involve, at its mildest, the recurrence of stage 1 demarcation line or more severe features such as stage 3 and/or plus disease. The natural history of such reactivations may not follow through the typical stages of ROP either and may just show features of plus disease associated with fine neovascularisation [
34]. Such reactivation has been described more than 12 months after treatment with anti-VEGF (Bevacizumab) [
73]. In the RAINBOW study, 31% of infants receiving 0.2 mg of intravitreal Ranibizumab had one or more additional treatments during their follow-up with a median time to first retreatment of 55 days (range 29–111 days) [
64]. Ranibizumab may be associated with more rapid reactivation than bevacizumab or aflibercept [
74]. Reactivation of ROP following intravitreal anti-VEGF injection may also be seen more in more posterior aggressive diseases or in those injected at a younger post-menstrual age [
75]. Very low-dose bevacizumab for treatment of ROP, ranging from 0.002 to 0.016 mg, is also associated with earlier reactivations with a mean of 76.4 days to reactivation compared to low-dose bevacizumab, ranging from 0.031 to 0.25 mg, with a mean of 85.7 days to reactivation [
61].
Babies treated with anti-VEGF may never fully vascularise to the peripheral retina leaving a persistent area of avascular retina (PAR) that runs the risk of late reactivation of ROP. This potential for late reactivation in a persistent avascular zone leaves a number of management dilemmas. PAR occurs in untreated, spontaneously regressed ROP as well as in ROP treated with anti-VEGF. Long-term pathological changes within the PAR observed in adult patients with a history of ROP that never reached treatment criteria include atrophic retinal holes, retinal tears, both of which can lead to retinal detachment, tractional retinoschisis (splitting of the retinal layers), vascular anomalies including telangiectatic vessels, vascular leakage and neovascularisation [
76]. Similarly fundus fluorescein angiography studies carried out in eyes treated with a history of anti-VEGF injections with PAR demonstrate vascular anomalies such as vascular leakage, shunts, abnormal vessel branching and tangles [
77]. There is no consensus regarding the management of such cases in terms of treatment or follow-up regimes. Additionally, in the short to medium term, to continue to undertake regular, serial ROP examinations require parents or guardians to return with their baby for ongoing retinal examinations that may pose practical difficulties, particularly if they have to travel a significant distance. Furthermore, the examinations become increasingly stressful and difficult to carry out in infants with increasing age. Further research is required to answer the questions of whether and when to treat or not.