Bevacizumab (Avastin; Genentech Inc., San Francisco, California, USA) is a full anti-VEGF human monoclonal antibody approved in 2004 by the FDA for the treatment of metastatic colon cancer and used off label in the treatment of neovascular retinal diseases, such as retinopathy of prematurity (ROP) [
The first prospective case series was published in 2008 and proved neovascular regression after bevacizumab injection, in 17 of the 18 eyes included in the study [
BEAT-ROP (Bevacizumab Eliminates the Angiogenic Threat of Retinopathy of Prematurity) was the first prospective, controlled, randomized trial that investigated the effect of one dose of bevacizumab (0.625 mg) for stage 3+ ROP, without combining laser and bevacizumab in the same infant. The recurrence rate in infants following bevacizumab was significantly lower as compared to the laser recurrence rate, 4% versus 22%. Benefits of bevacizumab over laser were statistically significant in zone I ROP. Another important observation of the BEAT-ROP study was that the injection of bevacizumab allows continued peripheral vascularization into avascular retina. However, the BEAT-ROP study has some unresolved issues: safety, long-term outcomes, and appropriate dosage [
In a previous study, we reported an AP-ROP regression rate of 54.16%, following laser treatment [
We report our experience with intravitreal bevacizumab over a 55-month period for the treatment of ROP.
This is a retrospective case series.
The work was conducted in accordance with the Declaration of Helsinki (1964). All the intravitreal injections were carried out by the same ophthalmologist in the Departments of Neonatology and Ophthalmology belonging to the “Iuliu Hațieganu” University of Medicine and Pharmacy from Cluj-Napoca and in the Departments of Neonatology belonging to the “Saint Pantelimon” and “Polizu” Hospitals from Bucharest, Romania. The patients were enrolled after obtaining the informed consent from the parents/tutors. The study was approved by the Ethics Committee of the “Iuliu Hațieganu” University of Medicine and Pharmacy.
We analyzed retrospectively the files of all the consecutive infants with AP-ROP and zone I stage 3+ ROP who were treated by one intravitreal bevacizumab injection, from January 1, 2009, throughout July 31, 2013.
The treated premature infants came from a screening program that included the preterm newborns who met the following criteria: gestational age (GA) less than or equal to 33 weeks or birth weight (BW) less than or equal to 1500 grams. In the screening for ROP the premature infants with GA of more than 33 weeks and with BW of more than 1500 grams were also included, if other risk factors were associated: prolonged oxygen administration with saturation over 93%, repeated transfusions, sepsis, and necessity of more than 6 days of mechanical ventilation for the cardiorespiratory support. The first eye exam was performed at 4 weeks after birth in all the premature infants.
We classified our cases according to the ICROP revisited [
A dose of bevacizumab of 0.625 mg in 0.025 mL was injected in each eye from the study sample, according to the following protocol: the pupils were dilated with a mixture of tropicamide 0.5% and phenylephrine 0.5%. Anesthesia was achieved with 0.5% proparacaine hydrochloride administered topically, 3 times, every 2 minutes before the injection. Every eye was prepared in a sterile manner using 5% povidone iodine which was also instilled in the eye 3 minutes prior to the injection. A nurse held the infant’s head during the procedure and 0.025 mL of bevacizumab (0.625 mg) was injected in pars plicata, 1.5–1.75 mm away from the limbus, with a 30 G needle, perpendicularly on the globe initially and then slightly directed toward the center of the eyeball. After the injection, the patients received topical tobramycin, 5 times/day, for 3 days. The patients were reexamined the next day and then every week to monitor the regression of the disease. The follow-up continued for 60 weeks, every 2 weeks initially and then every 3 weeks, until full vascularization of the retina was observed. Full vascularization was defined as follows: vascularization as far as it would develop without an active component or clinically significant tractional elements. All the patients were followed by the pediatricians for long-term systemic complications. All the intravitreal injections were carried out by the same ophthalmologist and the examinations were performed by three ophthalmologists experienced in ROP.
We considered the following to be signs of regression: the pupil dilation, the disappearance or decrease of retinal vessel tortuosity and neovascularization, and the growth of the normal retinal vessels toward the peripheral retina. The treatment failure of ROP was defined according to the following criteria: the persistence/reappearance of plus disease and of retinal neovascularization and the progression toward retinal detachment. In all these latter situations, conventional laser photocoagulation of the retina was carried out.
Pictures of the retina before and after bevacizumab injections were taken with a Ret Cam (Clarity Medical System, Pleasanton, California, USA).
The following data were recorded: gender, GA, BW, zone and stage of ROP, other ocular findings (persistence of fetal vasculature, vitreous hemorrhage, and undilated pupil), postnatal age (PNA) and postmenstrual age (PMA) at treatment, type of pregnancy (unique, multiple), complications, and follow-up period.
Visual acuity (VA) testing was performed at 60 weeks by a trained pediatric ophthalmologist using Teller acuity cards (Stereo Optical Company, Inc., Chicago, IL) at a test distance of 40 cm. VA data were compared with available data of normal VA ranges for full-term infants. We considered as normal a VA within one or two octaves or greater than two octaves of the lower limit of the normal range. In cases with ROP progression towards retinal detachment, VA was categorized as unrecordable.
We performed statistical analysis with the program IBM SPSS 20. Chi-square test was performed in order to verify the observed distributions in the contingency tables. If the theoretical distribution had less than 2 cases per cell we preferred the
Our study was performed on 74 eyes with zone I stage 3+ ROP and AP-ROP belonging to 37 premature infants, treated by intravitreally injected bevacizumab (Table
Demographic and clinical characteristics of the premature infants who received intravitreal bevacizumab as primary therapy for ROP.
Eyes (patients) | 74 (37) |
Male/female | 17/20 |
Unique/multiple pregnancy | 29/8 |
Mean gestational age (wk) | 28.89 |
Mean birth weight (g) | 1218.35 |
Mean injection time (postmenstrual age/postnatal age, wk) | 35.11/6.22 |
Number of eyes (%) with regression after bevacizumab injection alone | 63 (85.13) |
All of the infants had a history of supplemental oxygen and mechanical ventilation use. None of the infants had a history of laser prior to the bevacizumab injection. Within the study sample, 17 infants were males (45.94%), 20 were females (54.05%), 29 were coming from unique pregnancies (78.37%), and 8 were coming from multiple pregnancies (21.62%). The mean GA was
The ROP classification within the study sample is presented in Table
Classification of ROP in the eyes treated with intravitreal injections of bevacizumab.
ROP classification | Eyes (number) | Eyes (%) |
---|---|---|
AP-ROP |
52 | 70.27 |
Stage 3 zone I with “plus” | 22 | 29.72 |
Of the 37 prematures, 26 had bilateral aggressive posterior disease (AP-ROP) (70.27%) and 11 had bilateral zone I stage 3+ ROP (29.72%). We noted the persistence of fetal vasculature in 10 eyes (13.51%), undilated pupil in 7 eyes (9.45%), and vitreous hemorrhage in 4 eyes (5.71%), all within the AP-ROP group.
The most important data of each preterm newborn included in this study are illustrated in Table
Data of each premature infant treated by intravitreal injections of bevacizumab.
Case | Gender | GA (w) | BW (g) | ROP |
PNA (w) | PMA (w) | Outcome |
---|---|---|---|---|---|---|---|
1 | M | 28 | 900 | AP |
8 | 36 | Good/bad |
2 | F | 28 | 1000 | AP/AP | 5 | 33 | Good/good |
3 | F | 32 | 1800 | AP/AP | 4 | 36 | Good/good |
4 | M | 34 | 2300 | 3 I |
4 | 38 | Good/good |
5 | F | 27 | 995 | AP/AP | 7 | 34 | Good/good |
6 | F | 30 | 1400 | 3 I/3 I | 4 | 34 | Good/good |
7 | M | 30 | 1500 | 3 I/3 I | 4 | 34 | Good/good |
8 | F | 26 | 1100 | AP/AP | 5 | 31 | Good/good |
9 | F | 28 | 990 | AP/AP | 6 | 34 | Good/good |
10 | M | 30 | 1650 | 3 I/3 I | 7 | 37 | Good/good |
11 | F | 28 | 2050 | 3 I/3 I | 6 | 34 | Good/good |
12 | F | 27 | 1100 | AP/AP | 8 | 35 | Good/good |
13 | F | 29 | 1140 | AP/AP | 8 | 37 | Good/bad |
14 | M | 30 | 1200 | 3 I/3 I | 7 | 37 | Good/good |
15 | F | 27 | 970 | AP/AP | 6 | 33 | Good/good |
16 | F | 29 | 1070 | AP/AP | 6 | 35 | Good/good |
17 | F | 28 | 990 | AP/AP | 7 | 35 | Good/good |
18 | M | 28 | 1000 | AP/AP | 6 | 34 | Good/good |
19 | F | 31 | 1200 | 3 I/3 I | 5 | 36 | Good/good |
20 | F | 31 | 1100 | 3 I/3 I | 5 | 36 | Good/good |
21 | F | 30 | 1030 | AP/AP | 6 | 36 | Good/good |
22 | M | 28 | 990 | AP/AP | 7 | 35 | Good/good |
23 | M | 28 | 1100 | AP/AP | 7 | 35 | Good/good |
24 | M | 28 | 985 | AP/AP | 7 | 35 | Good/good |
25 | M | 31 | 1400 | AP/AP | 4 | 35 | Bad/bad |
26 | M | 32 | 1500 | 3 I/3 I | 5 | 37 | Good/good |
27 | M | 31 | 1240 | AP/AP | 5 | 36 | Good/good |
28 | F | 31 | 1499 | AP/AP | 7 | 38 | Bad/bad |
29 | M | 27 | 1010 | AP/AP | 7 | 34 | Good/good |
30 | F | 26 | 1100 | AP/AP | 8 | 34 | Bad/bad |
31 | F | 27 | 950 | AP/AP | 7 | 34 | Bad/bad |
32 | M | 27 | 980 | AP/AP | 5 | 32 | Good/good |
33 | M | 28 | 1550 | 3 I/3 I | 8 | 36 | Good/good |
34 | F | 27 | 800 | AP/AP | 10 | 37 | Good/good |
35 | M | 28 | 1390 | AP/AP | 7 | 35 | Good/bad |
36 | F | 31 | 1550 | 3 I/3 I | 6 | 37 | Good/good |
37 | M | 28 | 1050 | AP/AP | 6 | 34 | Good/good |
One week after the bevacizumab injection, ROP regressed in 63 eyes (85.13%) (Table
In all of the 11 eyes that failed to regress after one intravitreal injection of bevacizumab, laser photocoagulation was carried out at 7–10 days after the intravitreal bevacizumab, with good evolution in 9 of the 11 eyes. At 60 weeks, retina was attached completely in 72 eyes (97.29%) and totally detached (stage V) in 2 eyes (2.70%). None of the cases in this series progressed towards bilateral retinal detachment.
From the 72 eyes with good structural outcomes at 60 weeks, 68 had VA within normal limits (94.44%). All the 4 eyes with lower than normal VA come from the AP-ROP subgroup with laser addition.
We analyzed the following individual factors in correlation with the evolution after the intravitreal injection of bevacizumab: ROP type, gender, type of pregnancy, GA, and BW. Bevacizumab administered as one intravitreal injection was followed by ROP regression in all the 22 eyes with zone I stage 3+ ROP (100%) and in 41 of the 52 eyes with AP-ROP (78.84%) (Table
Within the group of girls (40 eyes), ROP regressed in 33 eyes (82.50%) and within the group of boys (34 eyes), ROP regressed in 30 eyes (88.23%). The difference between genders was not statistically significant (
Of the 37 pregnancies, 29 were unique and 8 were multiple. Within the unique pregnancies group (58 eyes), ROP regressed in 47 eyes (81.03%) and within the multiple pregnancies group (18 eyes), ROP regressed in all the 16 eyes (100.00%). The difference was not statistically significant (
An independent
An independent
We evaluated the moment of bevacizumab injections according to two parameters: PNA and PMA at treatment. PNA at the moment of bevacizumab injection varied on our series between 4 and 10 weeks (mean 6.22 ± 1 weeks) (Table
Figures
(a) AP-ROP prior to the bevacizumab injection: male infant, BW 1000 g, GA 28 w, treated with intravitreal bevacizumab at PMA of 34 weeks and PNA of 6 weeks. (b) Fundus photography 5 days after bevacizumab injection with venous dilatation and arteriolar tortuosity improvement in the posterior pole, showing plus disease regression.
The immediate good outcome of this case is revealed by the significant improvement of “plus” disease and the decrease of retinal vessel tortuosity.
Figure
(a) Zone I disease prior to the bevacizumab injection: female infant, BW 2050 g, GA 28 w, treated with intravitreal bevacizumab at PMA of 34 weeks and PNA of 6 weeks. (b) The same case as in (a), 8 weeks after the intravitreal injection of bevacizumab showing ROP regression.
We recorded no systemic complication subsequent to the intravitreal administration of bevacizumab in our series. No cataracts, endophthalmitis, or retinal detachments produced by the intraocular injections were identified in our series. The final anatomic outcome (after having added the laser photocoagulation of the retina) proved the bilateral attached retina in 35 cases (94.59%) and the unilateral attached retina in 2 cases (5.40%). No case ended up with bilateral retinal detachment.
The basis for the intravitreal administration of anti-VEGF agents as treatment for ROP is the evidence that the concentration of VEGF is increased in the vitreous of infants with ROP [
As opposed to the laser treatment, the intravitreal injection of bevacizumab is shorter, easier, accessible, and less expensive; it can be performed if the pupils are small and/or vitreous hemorrhage is associated and it is more efficient than laser in zone I ROP. It was not our goal to compare the results of bevacizumab with laser. However, we found that the primary intravitreal injection of 0.025 mL (0.625 mg) of bevacizumab was followed by the regression of ROP in 78.84% of the AP-ROP eyes. In the lasered AP-ROP cases, we found regression in 53.84% [
In our series, we used bevacizumab intravitreally in AP-ROP and in zone I stage 3+ ROP. In AP-ROP, our choice was motivated by our poor outcomes following laser treatment. As mentioned above, in our previous study, we registered regression rates of AP-ROP after laser treatment of 53.84% [
In zone I stage 3+ ROP, we used bevacizumab instead of laser, despite the good results of the laser treatment that we previously reported [
ROP is rare in infants with BW > 2000 g. On our series, we identified 2 infants with severe ROP, with BW over 2000 g. Also, 6 infants had BW over 1500 g, outside the screening criteria for ROP that are used in the USA. Nine of the 37 prematures had GA higher than 30 weeks, which is the limit for screening in the USA. This observation proves that, in each region, specific criteria for ROP screening must be applied. Older and bigger babies in developing and middle-income countries are known to develop significant ROP and possibly oxygen-induced retinopathy [
The exact therapeutic window during which intravitreal bevacizumab is effective is not precisely known. The most important decisional factors are the PMA at treatment and the degree of membrane formation [
In the neovascular retinal diseases of the adult (age-related macular degeneration and diabetic retinopathy), there is a continuous release of VEGF [
The only factor that influenced the outcomes of bevacizumab monotherapy in our series was the ROP type. ROP regressed in all the eyes with zone I stage 3+ ROP (100%) and in 78.84% of the eyes with AP-ROP after one intravitreal bevacizumab injection (
Recent studies proved that VEGF is not the only growth factor upregulated in the eye and therefore its inhibition may not induce the ROP regression in all cases. Other growth factors involved in ROP pathogeny are insulin-like growth factor, angiopoietin-1, and angiopoietin-2 [
Within the group of eyes with good structural outcomes (72 eyes), 94.44% had normal VA at 60 weeks. This observation allows us to affirm that if the structural result was successful, the potential for normal VA development was high in our series.
All the patients in the study sample were followed by pediatricians for systemic side effects related to bevacizumab, regarding the development of the brain, lungs, kidneys, and skeleton, and no abnormalities were reported. No local side effects, such as cataract, endophthalmitis, vitreous hemorrhage, and retinal detachment related to the intraocular injection, were identified in our series. ROP progression in 11 AP-ROP eyes was probably not due to the bevacizumab injection but to its inefficacy.
Still, there are concerns regarding the long-term effects of bevacizumab, the probability of the substance to leave the eye and the delayed-onset retinal detachment [
In the eye, VEGF is not only an angiogenic factor, but also a neural survival one. Therefore, the suppression of the normal development of neural retinal components by the anti-VEGF therapy is suspected [
In the eyes with ROP, there is the breakdown of the blood-retinal barrier. As a consequence, the anti-VEGF substances might leave the vitreous and get into the systemic circulation, reducing the serum VEGF level [
Even if there has been no report on the possible systemic side effects of intravitreally administered bevacizumab, we do not know for sure the long-term side effects on a developing child. Larger and longer studies are needed, in order to fully address the systemic toxicity issues.
Bevacizumab changes the natural evolution of ROP; despite the initial resolution there are chances of recurrence that we must be aware of. The delayed-onset retinal detachment, 4–4.5 months after the injection, is explained by the incomplete regression of the fibrovascular proliferation that develops slowly into a tractional complex [
A recent study documented significant vascular and macular abnormalities at the age of 9 months, in the eyes that received intravitreal bevacizumab for ROP [
The limitations of our study are its retrospective nature and the lack of a concomitant control group. On the ground of our results, currently, bevacizumab is our treatment of choice in AP-ROP and zone I disease. We follow each patient for 15 months after the injection, in order to identify late onset retinal detachment. We still have questions to be answered, regarding the optimal dosing, timing, avoidance of unnecessary treatment, long-term effects, and safety.
ROP regression rate after one intravitreal bevacizumab injection was 85.13%. Our study proves that 100% of patients with zone I stage 3+ ROP and 78.84% of patients with AP-ROP regressed with a single injection of bevacizumab, without ocular or systemic complications. Eyes with zone I stage 3+ ROP were more likely to have regression of ROP with a single injection of bevacizumab than those with AP-ROP.
The authors declare that there is no conflict of interests regarding the publication of this paper.
This paper was performed within the frame of the Program “Partnerships in priority domains, PN II, with the support of the National Education Ministry, UEFISCDI, Project no. 134.”