Successful regression of newly formed corneal neovascularization by subconjunctival injection of bevacizumab in patients with chemical burns

Abstract

Purpose: To investigate the effect and timing of subconjunctival bevacizumab injection on inhibiting corneal neovascularization (CorNV) in patients after chemical burns.

Methods: Patients with CorNV secondary to chemical burns were involved. Two subconjunctival injections of bevacizumab (2.5 mg/0.1 mL per involved quadrant) with an interval of 4 weeks were administered, and followed up a year. The area occupied by neovascular vessels (NA), accumulative neovascular length (NL), mean neovascular diameter (ND), best-corrected visual acuity (BCVA) and intraocular pressure (IOP) were evaluated. Complication was also recorded.

Results: Eleven patients with CorNV were involved. Eight patients had a history of surgery (four had amniotic grafts, one had keratoplasty, and three had amniotic grafts and keratoplasty). Decreasing in NA, NL, and ND were statistically significant at each time point compared to the baseline (p < 0.01). CorNV that developed within 1 month was considerably regressed, and vessels with fibrovascular membranes were found to be narrower and shorter than pretreatment. BCVA improved in five patients (from one to five lines), remained unchanged in five patients, and decreased in one patient compared to pretreatment.

Conclusion: Subconjunctival bevacizumab injection has a particular potential for the regression of CorNV, especially newly formed within 1 month in patients after chemical burns.

Keywords: bevacizumab; chemical burns; corneal neovascularization; ocular surface reconstructions; wound healing.

Figure 1

Quantification of corneal neovascular vessels. Front-view corneal photographs (A) were analyzed using Adobe Photoshop and Image-Pro Plus 4.0. The total corneal area was identified and measured using Adobe Photoshop, the corneal vessels were manually delineated using the computer’s cursor (B), and the length and diameter of the vessels were determined using Image-Pro. NA (C): neovascular area, the area occupied by the neovascular vessels. NL (D): neovascular length, the approximate cumulative length of all vessels. ND (E): neovascular diameter, the mean diameter of all vessels. When determining NL and ND, only the longest branch and the widest branch were measured for one vessel tree.

Figure 2

Changes in corneal neovascularization (CorNV) of 11 patients during follow-up. The percent decrease in NA (A), NL (B), and ND (C) were described by three box plots. It could reflect the changes relative to the treatment effects at each time point along with the pointwise 95% confidence intervals. Changes in newly developed and more than one-month-old CorNV during follow-up (D–F). Box plots show the percent decreases and relative effects of the three parameters, respectively, at each follow-up time.

Figure 3

Sample Cases. Case 7 (A–C): Subconjunctival bevacizumab was administered to treat newly formed corneal vessels in the right eye of Case 7. A 19-year-old man with CorNV caused by an alkali burn (lime) had undergone anterior deep lamellar keratoplasty of the central cornea 20 days before presentation. The newly formed superficial vessels encroached onto the graft from the 3 to 9 o’ clock positions [blue arrows, (A)], and stromal vessels entered the interface of the lamellar graft and host native cornea [red arrows, (A)]. After the first injection, the new superficial corneal vessels were significantly narrower at four weeks [blue arrows, (B)], and only a small stromal vessel at 7 o’ clock remained at 12 months post-treatment [red arrow, (C)]. Case 10 (D–F): Subconjunctival bevacizumab was administered to treat newly developed CorNV in the right eye of Case 10. A 47-year-old man who had an alkali burn (lime) two months prior and amniotic membrane transplantation 11 days before presented with newly formed superficial CorNV (blue arrow) from the 5 to 11 o’ clock position and residual amniotic membrane (yellow arrow) on the superior cornea (D). The corneal vessels were considerably regressed (blue arrow) at 4 weeks (E) and 12 months (F) after treatment. The vessels in the fibrovascular membrane at the nasal-inferior cornea became narrower [red arrows, (E,F)]. Case 2 (G,H): Alleviation of corneal lipid deposition in Case 2 after bevacizumab injection. The right eye of a 48-year-old man with a history of acid burns (sulfuric) 22 months before showed stable corneal vessels with dense lipid deposition in the horizontal middle cornea [yellow arrow, (G)]. The iris could not be observed using a slit–lamp microscope. Six weeks after the subconjunctival bevacizumab injection, corneal lipid deposition was slightly alleviated [yellow arrow, (H)], and the iris was observed.